A model is still running

A model is still running or downloading. Leaving will stop the current operation. Stay on this page to let it finish.

Skip to content
LocalAI.toolsBeta

Join the waitlist

Get an email when the new apps and model runs launch.

No account needed.

ACE-Step 1.5

Generate 30–60-second instrumental music from an English description locally on desktop.

Desktop inferencePrivate · Runs on your computer
About this model

Generate 30–60-second instrumental music from an English description locally on desktop.

License
MIT
Model released
Jan 28, 2026
Source repository created
Jan 23, 2026
Model weights · FP32
6.3 GB

Best for

  • Instrumental music sketches and creative drafts
  • Local caption-conditioned music experimentation

Keep in mind

  • Music can omit requested instruments or structure; review output before use. CLAP checks are a narrow prompt-matching proxy, not a broad human listening evaluation.
  • Only caption-conditioned instrumental clips are exposed. Lyrics, vocals, reference audio, covers, editing and the optional planning language model are not enabled.
  • Descriptions are limited to 2000 characters and 256 formatted prompt tokens; ambiguous descriptions may be misinterpreted.
  • Duration is restricted to 30–60 seconds: a 10-second probe produced near-silence. A silent output raises an actionable error.
  • No content safety classifier is included; outputs may be unsuitable or resemble existing material.
  • FP32 CPU and Apple Metal are measured on this M3 Max only. FP16 MPS aborts with a native datatype assertion; other platforms and precisions are unverified.

Processed on your deviceYour input is processed locally in the desktop app. Model and runtime files download from external hosts.

Target platforms · verification required

  • macos · aarch64, x86_64
  • windows · x86_64

Recommended configuration

FP32 · mps (verified platform only)

Sources & technical details
Original model
https://huggingface.co/ACE-Step/Ace-Step1.5
Release verification
Verified
Release evidence
Original ACE-Step project explicitly announces Released ACE-Step v1.5 on 2026.01.28; the original Turbo checkpoint is present in the linked official v1.5 bundle. Repository creation and paper dates are separate.
Source updated
Feb 3, 2026
Native model files
https://huggingface.co/ACE-Step/Ace-Step1.5
Files updated
Feb 3, 2026
Artifact revision
19671f406d603126926c1b7e2adc169acbcade22

Desktop app

Sample result

ACE-Step 1.5 Turbo

Run ACE-Step 1.5 on your computer

Download the desktop app to load this model, provide input and view results locally.

Public desktop downloads are coming soon.

Device requirements

macOS Apple SiliconCPU / FP32Tested32.0 GiBRAM estimate
macOS Apple SiliconMPS / FP32Tested32.0 GiBRAM estimate
Windows Intel / AMDCPU / FP32Unverified32.0 GiBRAM estimate
macOS Intel / AMDCPU / FP32Unverified32.0 GiBRAM estimate
If your device falls short

MemoryAdvisory only: close other apps, reduce image/tile/token/batch settings or choose a smaller verified model/precision. CPU can need more RAM than GPU.

GPUIf the selected GPU runtime fails, existing CPU fallback may recover when enabled, but CPU has its own memory/runtime needs.

PlatformUse a compatible OS/architecture/runtime or another model. These estimates do not add hardware detection or block runs.

Benchmark evidence

Saved desktop measurements apply to the recorded hardware and runtime. Fallback cases report the device that actually ran; they do not prove GPU support on the requested device.

ACE-Step 1.5 Turbo

40 configurations tested
Passed38
Failed2
TestedNov 5, 2026
Runtimeace-step-1.5 19671f4

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedFastmacos · aarch646307 ms / 0 ms0.210 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646386 ms / 0 ms0.213 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646389 ms / 0 ms0.213 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646497 ms / 0 ms0.217 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646534 ms / 0 ms0.218 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646615 ms / 0 ms0.221 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646781 ms / 0 ms0.226 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646782 ms / 0 ms0.226 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646808 ms / 0 ms0.227 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646928 ms / 0 ms0.231 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646972 ms / 0 ms0.232 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647013 ms / 0 ms0.234 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647083 ms / 0 ms0.236 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647202 ms / 0 ms0.240 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647289 ms / 0 ms0.243 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647402 ms / 5994 ms0.247 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6414968 ms / 0 ms0.249 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647599 ms / 0 ms0.253 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647760 ms / 0 ms0.259 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647889 ms / 0 ms0.263 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647953 ms / 0 ms0.265 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648132 ms / 0 ms0.271 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648309 ms / 2529 ms0.277 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648732 ms / 2513 ms0.291 RTF
FP32cpuPassedFastmacos · aarch6416543 ms / 0 ms0.551 RTF
FP32cpuPassedFastmacos · aarch6416716 ms / 0 ms0.557 RTF
FP32cpuPassedFastmacos · aarch6416727 ms / 0 ms0.558 RTF
FP32cpuPassedFastmacos · aarch6416744 ms / 0 ms0.558 RTF
FP32cpuPassedFastmacos · aarch6433651 ms / 0 ms0.561 RTF
FP32cpuPassedFastmacos · aarch6416962 ms / 0 ms0.565 RTF
FP32cpuPassedFastmacos · aarch6416988 ms / 1265 ms0.566 RTF
FP32cpuPassedFastmacos · aarch6417268 ms / 0 ms0.576 RTF
FP32cpuPassedFastmacos · aarch6417273 ms / 0 ms0.576 RTF
FP32cpuPassedFastmacos · aarch6417561 ms / 4374 ms0.585 RTF
FP32cpuPassedFastmacos · aarch6417645 ms / 1388 ms0.588 RTF
FP32cpuPassedFastmacos · aarch6417767 ms / 0 ms0.592 RTF
FP32cpuPassedFastmacos · aarch6419065 ms / 0 ms0.636 RTF
FP32cpuPassedFastmacos · aarch6420666 ms / 0 ms0.689 RTF
FP32cpuFailed—macos · aarch6416642 ms / 0 ms0.555 RTF
FP32mpsFailed—macos · aarch646648 ms / 0 ms0.222 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-05T00:20:42.586148+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "PyTorch 2.10.0 / Transformers 4.57.6 / Diffusers 0.37.0 / reviewed ACE-Step 19671f4", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 19671f4
Artifact revision
19671f406d603126926c1b7e2adc169acbcade22
Fixture
{
  "file": "model-samples/desktop/ace-step/ace-step-piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/ace-step-piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1264.7296669892967,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16988.358750008047,
      "durationMs": 17012.00595899718,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7659151446861099,
      "realTimeFactor": 0.5662786250002683,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 18276.735625986475,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": false
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16962.342167040333,
      "durationMs": 16984.629167011008,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.768623678532629,
      "realTimeFactor": 0.5654114055680111,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 16984.629167011008,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991557598114014
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0003233635798096657
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003041888412553817
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001288434286834672
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00007107122655725107
        },
        {
          "label": "White noise and static.",
          "score": 0.00001669329867581837
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16727.304166997783,
      "durationMs": 16750.93087501591,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7934748899460229,
      "realTimeFactor": 0.5575768055665927,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06830237060785294,
      "clippedFraction": 0,
      "rawPeak": 1.1416207551956177,
      "rawClippedFraction": 0.000028125,
      "outputGain": 0.858428681801669,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 16750.93087501591,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.7450898885726929
        },
        {
          "label": "Solo piano music.",
          "score": 0.25165510177612305
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.002908745314925909
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002906020963564515
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00002954627234430518
        },
        {
          "label": "White noise and static.",
          "score": 0.000026143041395698674
        }
      ],
      "expected": "Solo acoustic guitar music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 17268.13908404438,
      "durationMs": 17294.63466699235,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7373035886489796,
      "realTimeFactor": 0.5756046361348126,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11547164618968964,
      "clippedFraction": 0,
      "rawPeak": 1.0159716606140137,
      "rawClippedFraction": 0.000006597222222222222,
      "outputGain": 0.9645938346427165,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 17294.63466699235,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "fail",
    "quality": {
      "passed": false,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.7860839366912842
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.21041683852672577
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0028349694330245256
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005501423147507012
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009836136450758204
        },
        {
          "label": "White noise and static.",
          "score": 0.00001575487658556085
        }
      ],
      "expected": "Solo acoustic guitar music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16642.379625001922,
      "durationMs": 16668.88062498765,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.8026268283732012,
      "realTimeFactor": 0.5547459875000641,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10291688144207001,
      "clippedFraction": 0,
      "rawPeak": 1.145395278930664,
      "rawClippedFraction": 0.0001277777777777778,
      "outputGain": 0.8555998248176154,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 16668.88062498765,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9661152958869934
        },
        {
          "label": "Solo piano music.",
          "score": 0.023513494059443474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.009294203482568264
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0004468896659091115
        },
        {
          "label": "White noise and static.",
          "score": 0.0003425213217269629
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00028763266163878143
        }
      ],
      "expected": "Electronic dance music with drums and synthesizers."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 17273.22070801165,
      "durationMs": 17300.110000011045,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7367924897807523,
      "realTimeFactor": 0.5757740236003883,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.09996964037418365,
      "clippedFraction": 0,
      "rawPeak": 2.1385786533355713,
      "rawClippedFraction": 0.000890625,
      "outputGain": 0.45824828489309016,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 17300.110000011045,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9783398509025574
        },
        {
          "label": "Solo piano music.",
          "score": 0.013563388958573341
        },
        {
          "label": "White noise and static.",
          "score": 0.0036051536444574594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0024097184650599957
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012954809935763478
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0007863707724027336
        }
      ],
      "expected": "Electronic dance music with drums and synthesizers."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16716.464083991013,
      "durationMs": 16744.172624952625,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.794637900052699,
      "realTimeFactor": 0.5572154694663671,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.09050016850233078,
      "clippedFraction": 0,
      "rawPeak": 2.2191367149353027,
      "rawClippedFraction": 0.005079861111111111,
      "outputGain": 0.4416131702947248,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 16744.172624952625,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 17766.9870830141,
      "durationMs": 17794.785291014705,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.688524895066824,
      "realTimeFactor": 0.5922329027671367,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 17794.785291014705,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 19065.305084048305,
      "durationMs": 19094.50841700891,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.573538942479374,
      "realTimeFactor": 0.6355101694682769,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 19094.50841700891,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 20665.799957991112,
      "durationMs": 20699.22258297447,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.4516737828191117,
      "realTimeFactor": 0.6888599985997037,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 20699.22258297447,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16743.892207974568,
      "durationMs": 16771.381500002462,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7916981086220791,
      "realTimeFactor": 0.558129740265819,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 16771.381500002462,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16542.732500005513,
      "durationMs": 16571.254874987062,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.8134851663707916,
      "realTimeFactor": 0.5514244166668505,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 16571.254874987062,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.967025637626648
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.016082584857940674
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.008739323355257511
        },
        {
          "label": "White noise and static.",
          "score": 0.0035762106999754906
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0031910811085253954
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0013850389514118433
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33651.02708298946,
      "durationMs": 33705.419999954756,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 1.7830064993864603,
      "realTimeFactor": 0.5608504513831576,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0363805815577507,
      "clippedFraction": 0,
      "rawPeak": 0.8011957406997681,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 33705.419999954756,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4374.114208039828,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 17560.682584007736,
      "durationMs": 17587.901709019206,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7083618393809232,
      "realTimeFactor": 0.5853560861335912,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 21962.015917059034,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": false
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9990785121917725
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000616762728895992
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015724649711046368
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010030900011770427
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000041727656935108826
        },
        {
          "label": "White noise and static.",
          "score": 0.000005461812179419212
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1387.8256669850089,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 17645.41462500347,
      "durationMs": 17674.102000019047,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.70015840588354,
      "realTimeFactor": 0.5881804875001156,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07040376216173172,
      "clippedFraction": 0,
      "rawPeak": 0.724529504776001,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "driverAllocatedMemoryBytes": null,
      "totalMs": 19061.927667004056,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": false
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5993.903082970064,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7401.7801659647375,
      "durationMs": 7432.199499977287,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.053079033331415,
      "realTimeFactor": 0.24672600553215793,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13938393088,
      "totalMs": 13426.102582947351,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": false
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6807.563000009395,
      "durationMs": 6834.337500040419,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.406863366517298,
      "realTimeFactor": 0.22691876666697983,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13938393088,
      "totalMs": 6834.337500040419,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.999160885810852
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00032731646206229925
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00028546462999656796
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00013199289969634265
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00007611326145706698
        },
        {
          "label": "White noise and static.",
          "score": 0.000018223767256131396
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6385.64466603566,
      "durationMs": 6416.825000022072,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.698037797117926,
      "realTimeFactor": 0.21285482220118865,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.0683024451136589,
      "clippedFraction": 0,
      "rawPeak": 1.1416208744049072,
      "rawClippedFraction": 0.000028125,
      "outputGain": 0.8584285921636153,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13938393088,
      "totalMs": 6416.825000022072,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.7481545209884644
        },
        {
          "label": "Solo piano music.",
          "score": 0.24869714677333832
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0028060502372682095
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002875539066735655
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000028963773729628883
        },
        {
          "label": "White noise and static.",
          "score": 0.00002564942406024784
        }
      ],
      "expected": "Solo acoustic guitar music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6615.042541990988,
      "durationMs": 6646.783249976579,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.535118226310097,
      "realTimeFactor": 0.22050141806636628,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11547203361988068,
      "clippedFraction": 0,
      "rawPeak": 1.0159692764282227,
      "rawClippedFraction": 0.000006597222222222222,
      "outputGain": 0.9645960982652177,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13944389632,
      "totalMs": 6646.783249976579,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "fail",
    "quality": {
      "passed": false,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.7852321863174438
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.2112632840871811
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.002828286262229085
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005609946674667299
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009921796299749985
        },
        {
          "label": "White noise and static.",
          "score": 0.000016062400391092524
        }
      ],
      "expected": "Solo acoustic guitar music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6647.638583031949,
      "durationMs": 6673.019749985542,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.512880720768242,
      "realTimeFactor": 0.22158795276773163,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10291754454374313,
      "clippedFraction": 0,
      "rawPeak": 1.1453742980957031,
      "rawClippedFraction": 0.0001277777777777778,
      "outputGain": 0.8556154975970265,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13944389632,
      "totalMs": 6673.019749985542,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9652296304702759
        },
        {
          "label": "Solo piano music.",
          "score": 0.024201689288020134
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.009481414221227169
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00045314227463677526
        },
        {
          "label": "White noise and static.",
          "score": 0.00034541054628789425
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00028873924748040736
        }
      ],
      "expected": "Electronic dance music with drums and synthesizers."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6533.784041996114,
      "durationMs": 6560.647000034805,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.591519984005288,
      "realTimeFactor": 0.21779280139987048,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09996932744979858,
      "clippedFraction": 0,
      "rawPeak": 2.1385841369628906,
      "rawClippedFraction": 0.000890625,
      "outputGain": 0.4582471098807207,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 6560.647000034805,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9782841801643372
        },
        {
          "label": "Solo piano music.",
          "score": 0.013598219491541386
        },
        {
          "label": "White noise and static.",
          "score": 0.0036094204988330603
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0024211250711232424
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0013034361181780696
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0007836279692128301
        }
      ],
      "expected": "Electronic dance music with drums and synthesizers."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6388.98008299293,
      "durationMs": 6415.683749946766,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.69558514978911,
      "realTimeFactor": 0.212966002766431,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09049992263317108,
      "clippedFraction": 0,
      "rawPeak": 2.2191357612609863,
      "rawClippedFraction": 0.005079513888888889,
      "outputGain": 0.4416133600781286,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 6415.683749946766,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6497.256832954008,
      "durationMs": 6522.164499969222,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.617333248678175,
      "realTimeFactor": 0.2165752277651336,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 6522.164499969222,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6307.0428330102,
      "durationMs": 6332.528041966725,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.756587325360167,
      "realTimeFactor": 0.21023476110034,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 6332.528041966725,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6781.388750008773,
      "durationMs": 6813.355959020555,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.4238726175315035,
      "realTimeFactor": 0.2260462916669591,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 6813.355959020555,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6928.124957950786,
      "durationMs": 6952.222833002452,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.330175939677834,
      "realTimeFactor": 0.23093749859835952,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 6952.222833002452,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7083.24966597138,
      "durationMs": 7110.417582967784,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.235344144951281,
      "realTimeFactor": 0.23610832219904598,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951303680,
      "totalMs": 7110.417582967784,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9670519232749939
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.016035720705986023
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00866205245256424
        },
        {
          "label": "White noise and static.",
          "score": 0.003635894972831011
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0032095443457365036
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0014047707663848996
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 14967.76933298679,
      "durationMs": 15050.446624984033,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 4.008613352142508,
      "realTimeFactor": 0.2494628222164465,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.03638068214058876,
      "clippedFraction": 0,
      "rawPeak": 0.8011802434921265,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026651136,
      "totalMs": 15050.446624984033,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6971.904541947879,
      "durationMs": 7012.649416981731,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.302984904554976,
      "realTimeFactor": 0.2323968180649293,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037136896,
      "totalMs": 7012.649416981731,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6782.135917048436,
      "durationMs": 6808.375917025842,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.423385253101195,
      "realTimeFactor": 0.22607119723494787,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037136896,
      "totalMs": 6808.375917025842,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7013.474374951329,
      "durationMs": 7038.341665989719,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.277480517665424,
      "realTimeFactor": 0.2337824791650443,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037136896,
      "totalMs": 7038.341665989719,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7201.877291023266,
      "durationMs": 7249.6114580426365,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.165580554585859,
      "realTimeFactor": 0.2400625763674422,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037136896,
      "totalMs": 7249.6114580426365,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7599.431999959052,
      "durationMs": 7624.155625002459,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.94766345697437,
      "realTimeFactor": 0.25331439999863503,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037136896,
      "totalMs": 7624.155625002459,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7289.4847500137985,
      "durationMs": 7317.26099998923,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.115517218133039,
      "realTimeFactor": 0.24298282500045995,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037136896,
      "totalMs": 7317.26099998923,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7759.860875026789,
      "durationMs": 7789.719958032947,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.8660486938042475,
      "realTimeFactor": 0.2586620291675596,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026651136,
      "totalMs": 7789.719958032947,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8132.006791012827,
      "durationMs": 8160.456707992125,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.689126284689631,
      "realTimeFactor": 0.2710668930337609,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026651136,
      "totalMs": 8160.456707992125,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7889.440040977206,
      "durationMs": 7931.555750023108,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.802551238640775,
      "realTimeFactor": 0.2629813346992402,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026651136,
      "totalMs": 7931.555750023108,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 0,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7952.997917018365,
      "durationMs": 7982.7512499759905,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.7721624364824695,
      "realTimeFactor": 0.2650999305672788,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026651136,
      "totalMs": 7982.7512499759905,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": true
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2529.1480000014417,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8308.823750005104,
      "durationMs": 8339.41579196835,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.610619373167179,
      "realTimeFactor": 0.2769607916668368,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13927907328,
      "totalMs": 10868.563791969791,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": false
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "CLAP top-1 instrument/style among six fixed music/speech/noise labels plus signal validity",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9991286396980286
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0005704713403247297
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00015609954425599426
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009526189387543127
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000044119344238424674
        },
        {
          "label": "White noise and static.",
          "score": 0.000005440802851808257
        }
      ],
      "expected": "Solo piano music."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "19671f4",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2513.3815829758532,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8731.851125019602,
      "durationMs": 8763.295292039402,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.435697605292446,
      "realTimeFactor": 0.2910617041673201,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07040384411811829,
      "clippedFraction": 0,
      "rawPeak": 0.7245202660560608,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13927907328,
      "totalMs": 11276.676875015255,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "19671f406d603126926c1b7e2adc169acbcade22",
      "fallback": false,
      "grade": "fast",
      "initializationReused": false
    }
  }
]

ACE-Step 1.5 Base

52 configurations tested
Passed52
Failed0
TestedNov 7, 2026
Runtimeace-step-1.5 e432212fec32b8965a14ffa57ae653438d6abd14

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedFastmacos · aarch6425914 ms / 6515 ms0.864 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6425916 ms / 6515 ms0.864 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6427058 ms / 6515 ms0.902 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6427657 ms / 6515 ms0.922 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431785 ms / 6515 ms1.060 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6438525 ms / 6515 ms1.284 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6441405 ms / 6515 ms1.380 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432163 ms / 6515 ms1.072 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6436606 ms / 6515 ms1.220 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431519 ms / 6515 ms1.051 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431965 ms / 6515 ms1.066 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6434297 ms / 6515 ms1.143 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6469915 ms / 6515 ms1.165 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6436381 ms / 6515 ms1.213 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433703 ms / 6515 ms1.123 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432312 ms / 6515 ms1.077 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432455 ms / 6515 ms1.082 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6434034 ms / 6515 ms1.134 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6435905 ms / 6515 ms1.197 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431314 ms / 6515 ms1.044 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430976 ms / 6515 ms1.033 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431116 ms / 6515 ms1.037 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430944 ms / 6515 ms1.031 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431768 ms / 6014 ms1.059 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431027 ms / 2318 ms1.034 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433525 ms / 2233 ms1.118 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6435316 ms / 2233 ms1.177 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430627 ms / 2233 ms1.021 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433507 ms / 2233 ms1.117 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433312 ms / 2233 ms1.110 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432244 ms / 2233 ms1.075 RTF
FP32cpuPassedNot measuredmacos · aarch6467791 ms / 4868 ms2.260 RTF
FP32cpuPassedNot measuredmacos · aarch6468485 ms / 4868 ms2.283 RTF
FP32cpuPassedNot measuredmacos · aarch6469078 ms / 4868 ms2.303 RTF
FP32cpuPassedNot measuredmacos · aarch6470772 ms / 4868 ms2.359 RTF
FP32cpuPassedNot measuredmacos · aarch6469483 ms / 4868 ms2.316 RTF
FP32cpuPassedNot measuredmacos · aarch6469615 ms / 4868 ms2.321 RTF
FP32cpuPassedNot measuredmacos · aarch6470257 ms / 4868 ms2.342 RTF
FP32cpuPassedNot measuredmacos · aarch6470469 ms / 4868 ms2.349 RTF
FP32cpuPassedNot measuredmacos · aarch6469414 ms / 4868 ms2.314 RTF
FP32cpuPassedNot measuredmacos · aarch6470121 ms / 4868 ms2.337 RTF
FP32cpuPassedNot measuredmacos · aarch6469761 ms / 4868 ms2.325 RTF
FP32cpuPassedNot measuredmacos · aarch6466462 ms / 4868 ms2.215 RTF
FP32cpuPassedNot measuredmacos · aarch64126903 ms / 4868 ms2.115 RTF
FP32cpuPassedNot measuredmacos · aarch6465629 ms / 4936 ms2.188 RTF
FP32cpuPassedNot measuredmacos · aarch6466923 ms / 3048 ms2.231 RTF
FP32cpuPassedNot measuredmacos · aarch6474013 ms / 4890 ms2.467 RTF
FP32cpuPassedNot measuredmacos · aarch6475974 ms / 4890 ms2.532 RTF
FP32cpuPassedNot measuredmacos · aarch6472674 ms / 4890 ms2.422 RTF
FP32cpuPassedNot measuredmacos · aarch6471393 ms / 4890 ms2.380 RTF
FP32cpuPassedNot measuredmacos · aarch6471481 ms / 4890 ms2.383 RTF
FP32cpuPassedNot measuredmacos · aarch6471665 ms / 4890 ms2.389 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-07T19:07:24.618480+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Torch2.10.0/Transformers4.57.6/Diffusers0.37.0/exact reviewed ACEe432212fec32b8965a14ffa57ae653438d6abd14", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 e432212fec32b8965a14ffa57ae653438d6abd14
Artifact revision
e432212fec32b8965a14ffa57ae653438d6abd14
Fixture
{
  "file": "model-samples/desktop/ace-step-variants/base/piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 67790.56216600293,
      "durationMs": 67817.53212500189,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.44253947808453264,
      "realTimeFactor": 2.259685405533431,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 72685.10970799252,
      "postInferenceRssBytes": 14679998464,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 68484.7255420027,
      "durationMs": 68511.97108300403,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.438053883732082,
      "realTimeFactor": 2.2828241847334234,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 73379.54866599466,
      "postInferenceRssBytes": 14815084544,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9971644282341003
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0018849305342882872
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.000402605626732111
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0003741422842722386
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00013731633953284472
        },
        {
          "label": "White noise and static.",
          "score": 0.000036549041396938264
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69077.90324999951,
      "durationMs": 69100.12458398705,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4342922785514659,
      "realTimeFactor": 2.3025967749999836,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1323506236076355,
      "clippedFraction": 0,
      "rawPeak": 1.0705114603042603,
      "rawClippedFraction": 0.00002777777777777778,
      "outputGain": 0.9154502649803159,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 73967.70216697769,
      "postInferenceRssBytes": 14817902592,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9932020306587219
        },
        {
          "label": "Solo piano music.",
          "score": 0.006673825439065695
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00009804911678656936
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000014299890608526766
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000009109349775826558
        },
        {
          "label": "White noise and static.",
          "score": 0.00000272306192528049
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70771.55545901041,
      "durationMs": 70798.19466700428,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.42389911886810866,
      "realTimeFactor": 2.3590518486336807,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.12502144277095795,
      "clippedFraction": 0,
      "rawPeak": 0.9234083890914917,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75665.77224999492,
      "postInferenceRssBytes": 14828634112,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9867968559265137
        },
        {
          "label": "Solo piano music.",
          "score": 0.009182698093354702
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003617257811129093
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000375877832993865
        },
        {
          "label": "White noise and static.",
          "score": 0.000016890766346477903
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000010379467312304769
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69483.18799999834,
      "durationMs": 69505.04249999358,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.43175911847914517,
      "realTimeFactor": 2.316106266666611,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11256572604179382,
      "clippedFraction": 0,
      "rawPeak": 1.29166841506958,
      "rawClippedFraction": 0.00003506944444444444,
      "outputGain": 0.7587086504296143,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74372.62008298421,
      "postInferenceRssBytes": 14834991104,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9749001264572144
        },
        {
          "label": "Solo piano music.",
          "score": 0.012762777507305145
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.01196735817939043
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002169208601117134
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009812063217395917
        },
        {
          "label": "White noise and static.",
          "score": 0.00005472695193020627
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69615.01391700585,
      "durationMs": 69641.18183300889,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4309415212609973,
      "realTimeFactor": 2.320500463900195,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.2153574377298355,
      "clippedFraction": 0,
      "rawPeak": 1.4303689002990723,
      "rawClippedFraction": 0.006738541666666667,
      "outputGain": 0.6851379387478952,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74508.75941599952,
      "postInferenceRssBytes": 14838087680,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9994571805000305
        },
        {
          "label": "Solo piano music.",
          "score": 0.00017011506133712828
        },
        {
          "label": "White noise and static.",
          "score": 0.00016079236229415983
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00011147005716338754
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000058896646805806085
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004165383506915532
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70256.96645800781,
      "durationMs": 70284.97920800874,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4270039187918942,
      "realTimeFactor": 2.3418988819335937,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10896895080804825,
      "clippedFraction": 0,
      "rawPeak": 1.418222427368164,
      "rawClippedFraction": 0.00012152777777777777,
      "outputGain": 0.6910058542922735,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75152.55679099937,
      "postInferenceRssBytes": 14848966656,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70469.14800000377,
      "durationMs": 70491.93025000568,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.42571821643137214,
      "realTimeFactor": 2.3489716000001257,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75359.50783299631,
      "postInferenceRssBytes": 14852374528,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69413.9221249934,
      "durationMs": 69437.55420799425,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4321899567348911,
      "realTimeFactor": 2.3137974041664466,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74305.13179098489,
      "postInferenceRssBytes": 14858174464,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70120.64204200578,
      "durationMs": 70148.45791700645,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4278340746228264,
      "realTimeFactor": 2.337354734733526,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75016.03549999709,
      "postInferenceRssBytes": 14861238272,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69761.09874999383,
      "durationMs": 69784.66350000235,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.43003909825893694,
      "realTimeFactor": 2.3253699583331278,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74652.24108299299,
      "postInferenceRssBytes": 14872150016,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 66461.81054199405,
      "durationMs": 66486.47608299507,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.451387040999198,
      "realTimeFactor": 2.2153936847331352,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 71354.0536659857,
      "postInferenceRssBytes": 14875557888,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.8537506461143494
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.07565592974424362
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.06975056976079941
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000673580274451524
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00014671109966002405
        },
        {
          "label": "White noise and static.",
          "score": 0.000022657590307062492
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4867.577582990634,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 126902.54354100034,
      "durationMs": 126953.03804099967,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.47280376205079677,
      "realTimeFactor": 2.1150423923500057,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.09151027351617813,
      "clippedFraction": 0,
      "rawPeak": 0.9906056523323059,
      "rawClippedFraction": 0,
      "outputGain": 0.9892937696172682,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 131820.6156239903,
      "postInferenceRssBytes": 14986575872,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4935.926958001801,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 65629.46525000734,
      "durationMs": 65655.02820900292,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.45711175438834833,
      "realTimeFactor": 2.1876488416669115,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 70590.95516700472,
      "postInferenceRssBytes": 14684307456,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9933497309684753
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004092907998710871
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002247400349006057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018087781791109592
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011052590707549825
        },
        {
          "label": "White noise and static.",
          "score": 0.000018560433090897277
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 3047.835125005804,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 66922.54975000105,
      "durationMs": 66947.52658299694,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.44827939329970806,
      "realTimeFactor": 2.2307516583333684,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1287851780653,
      "clippedFraction": 0,
      "rawPeak": 0.8788315653800964,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 69995.36170800275,
      "postInferenceRssBytes": 14680129536,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 27057.92362500506,
      "durationMs": 27090.530542001943,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1087325256649063,
      "realTimeFactor": 0.9019307875001686,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13986037760,
      "totalMs": 33605.260124997585,
      "postInferenceRssBytes": 2664349696,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25915.786707992083,
      "durationMs": 25945.0531249895,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1575955743897368,
      "realTimeFactor": 0.8638595569330695,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13986037760,
      "totalMs": 32459.78270798514,
      "postInferenceRssBytes": 2672541696,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9971534013748169
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0018903507152572274
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00040482808253727853
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00037561680073849857
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00013883356587029994
        },
        {
          "label": "White noise and static.",
          "score": 0.00003686495256260969
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25914.376041997457,
      "durationMs": 25937.36454199825,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1576585888613056,
      "realTimeFactor": 0.8638125347332486,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.13235032558441162,
      "clippedFraction": 0,
      "rawPeak": 1.0705124139785767,
      "rawClippedFraction": 0.00002777777777777778,
      "outputGain": 0.9154494494443218,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13986037760,
      "totalMs": 32452.09412499389,
      "postInferenceRssBytes": 2686894080,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9932686686515808
        },
        {
          "label": "Solo piano music.",
          "score": 0.006608873140066862
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0000965945073403418
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000014078155800234526
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000008995741154649295
        },
        {
          "label": "White noise and static.",
          "score": 0.000002684836999833351
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 27656.593374995282,
      "durationMs": 27682.230416001403,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0847322948719862,
      "realTimeFactor": 0.921886445833176,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.12502115964889526,
      "clippedFraction": 0,
      "rawPeak": 0.9234222769737244,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13996425216,
      "totalMs": 34196.959998997045,
      "postInferenceRssBytes": 2707718144,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9874255061149597
        },
        {
          "label": "Solo piano music.",
          "score": 0.00892917811870575
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003283228026703
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00033693527802824974
        },
        {
          "label": "White noise and static.",
          "score": 0.00001527873973827809
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000009831241186475381
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31785.100625013,
      "durationMs": 31833.249209012138,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9438384466334446,
      "realTimeFactor": 1.0595033541671,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11268553137779236,
      "clippedFraction": 0,
      "rawPeak": 1.2905044555664062,
      "rawClippedFraction": 0.00003506944444444444,
      "outputGain": 0.7593929612353605,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13996425216,
      "totalMs": 38347.97879200778,
      "postInferenceRssBytes": 2710880256,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9749149680137634
        },
        {
          "label": "Solo piano music.",
          "score": 0.012731825932860374
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.011987782083451748
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00021507948986254632
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009659461647970602
        },
        {
          "label": "White noise and static.",
          "score": 0.00005372240411816165
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 38525.09816699603,
      "durationMs": 38566.20729199494,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.7787131357837946,
      "realTimeFactor": 1.2841699388998677,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.2153548002243042,
      "clippedFraction": 0,
      "rawPeak": 1.430384635925293,
      "rawClippedFraction": 0.006738194444444444,
      "outputGain": 0.6851304015622719,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14031372288,
      "totalMs": 45080.936874990584,
      "postInferenceRssBytes": 2735570944,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9994539618492126
        },
        {
          "label": "Solo piano music.",
          "score": 0.0001714356621960178
        },
        {
          "label": "White noise and static.",
          "score": 0.00016151093586813658
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00011167357297381386
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00005962288560112938
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000041841100028250366
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 41404.718541991315,
      "durationMs": 41428.132917004405,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.7245551003945354,
      "realTimeFactor": 1.3801572847330439,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1089387834072113,
      "clippedFraction": 0,
      "rawPeak": 1.4186195135116577,
      "rawClippedFraction": 0.00012152777777777777,
      "outputGain": 0.6908124346704517,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14031372288,
      "totalMs": 47942.86250000005,
      "postInferenceRssBytes": 773505024,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32162.553707996267,
      "durationMs": 32197.487042009016,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9327617536956149,
      "realTimeFactor": 1.0720851235998756,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 38712.21662500466,
      "postInferenceRssBytes": 756154368,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 36606.02804100199,
      "durationMs": 36652.89241599385,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8195371529081862,
      "realTimeFactor": 1.2202009347000664,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14035566592,
      "totalMs": 43167.62199898949,
      "postInferenceRssBytes": 765345792,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31519.180250004865,
      "durationMs": 31547.748500001035,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9518014035277891,
      "realTimeFactor": 1.050639341666829,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 38062.47808299668,
      "postInferenceRssBytes": 774815744,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31965.116749997833,
      "durationMs": 32002.17941700248,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9385230854819898,
      "realTimeFactor": 1.0655038916665944,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14035566592,
      "totalMs": 38516.90899999812,
      "postInferenceRssBytes": 783810560,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 34296.50604100607,
      "durationMs": 34333.18504100316,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8747246720739127,
      "realTimeFactor": 1.1432168680335357,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 40847.914623998804,
      "postInferenceRssBytes": 792772608,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.8530659675598145
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.07544038444757462
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.07064688950777054
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0006781641859561205
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00014590262435376644
        },
        {
          "label": "White noise and static.",
          "score": 0.00002261317968077492
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69915.20970800775,
      "durationMs": 69974.66320800595,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.8581823647612958,
      "realTimeFactor": 1.1652534951334625,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09150999784469604,
      "clippedFraction": 0,
      "rawPeak": 0.9906085133552551,
      "rawClippedFraction": 0,
      "outputGain": 0.9892909123914921,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 76489.39279100159,
      "postInferenceRssBytes": 988413952,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 36380.974458006676,
      "durationMs": 36407.38108300138,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8246068294467478,
      "realTimeFactor": 1.2126991486002225,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 42922.110665997025,
      "postInferenceRssBytes": 1000554496,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33703.47254099033,
      "durationMs": 33728.25604099489,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8901159951252456,
      "realTimeFactor": 1.1234490846996779,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 40242.985623990535,
      "postInferenceRssBytes": 1009713152,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32311.87829200644,
      "durationMs": 32335.209875003784,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9284511327037781,
      "realTimeFactor": 1.077062609733548,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 38849.939457999426,
      "postInferenceRssBytes": 1018839040,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32454.805749992374,
      "durationMs": 32480.448499991326,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9243623342286388,
      "realTimeFactor": 1.0818268583330792,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 38995.17808298697,
      "postInferenceRssBytes": 1027883008,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 34033.87420800573,
      "durationMs": 34061.77825000486,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8814747276977111,
      "realTimeFactor": 1.134462473600191,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 40576.507833000505,
      "postInferenceRssBytes": 1036894208,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 35904.89845900447,
      "durationMs": 35927.46525000257,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8355405888211445,
      "realTimeFactor": 1.1968299486334824,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 42442.19483299821,
      "postInferenceRssBytes": 1045872640,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31314.373040993814,
      "durationMs": 31338.022790994728,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9580265254146024,
      "realTimeFactor": 1.043812434699794,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 37852.75237399037,
      "postInferenceRssBytes": 1055571968,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30976.45650000777,
      "durationMs": 31000.776875007432,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9684774628754736,
      "realTimeFactor": 1.0325485500002591,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 37515.506458003074,
      "postInferenceRssBytes": 1064599552,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31116.018707994954,
      "durationMs": 31143.804374994943,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9641336278118318,
      "realTimeFactor": 1.037200623599832,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 37658.533957990585,
      "postInferenceRssBytes": 1073577984,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6514.729582995642,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30944.036584012792,
      "durationMs": 30968.45879200555,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9694921319831774,
      "realTimeFactor": 1.0314678861337598,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 37483.18837500119,
      "postInferenceRssBytes": 1082572800,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6014.01020900812,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31768.05037500162,
      "durationMs": 31804.652708000503,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9443450147512702,
      "realTimeFactor": 1.058935012500054,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14004912128,
      "totalMs": 37818.66291700862,
      "postInferenceRssBytes": 2671230976,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9936503767967224
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003825908759608865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0022106159012764692
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00018999638268724084
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00010408757952973247
        },
        {
          "label": "White noise and static.",
          "score": 0.00001904681630549021
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2318.057750002481,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31027.00712499791,
      "durationMs": 31057.031582997297,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9668995749135447,
      "realTimeFactor": 1.0342335708332635,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.128785640001297,
      "clippedFraction": 0,
      "rawPeak": 0.8788308501243591,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14004912128,
      "totalMs": 33375.08933299978,
      "postInferenceRssBytes": 2651291648,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9908740520477295
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.005266621243208647
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003125017276033759
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0004227797035127878
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00029041169909760356
        },
        {
          "label": "White noise and static.",
          "score": 0.00002117012263624929
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4890.193540995824,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 74012.51266599866,
      "durationMs": 74036.80520800117,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.40533686696171306,
      "realTimeFactor": 2.4670837555332885,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10620082914829254,
      "clippedFraction": 0,
      "rawPeak": 1.0328760147094727,
      "rawClippedFraction": 0.000005902777777777778,
      "outputGain": 0.9488070068851917,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4690.641791996313,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 78926.998748997,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9916185736656189
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.003967972006648779
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0034517336171120405
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0007751520606689155
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016957354091573507
        },
        {
          "label": "White noise and static.",
          "score": 0.000017110973203671165
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4890.193540995824,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 75973.97304199694,
      "durationMs": 76001.8475830002,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3948720699839744,
      "realTimeFactor": 2.5324657680665648,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10459727048873901,
      "clippedFraction": 0,
      "rawPeak": 1.1305488348007202,
      "rawClippedFraction": 0.000040625,
      "outputGain": 0.8668356198631109,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4361.699542001588,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 80892.04112399602,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9865185618400574
        },
        {
          "label": "Solo piano music.",
          "score": 0.013143234886229038
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021405605366453528
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00008529173646820709
        },
        {
          "label": "White noise and static.",
          "score": 0.00002136026341759134
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000017352360373479314
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4890.193540995824,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 72673.86770900339,
      "durationMs": 72701.12533400243,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4128031291815142,
      "realTimeFactor": 2.4224622569667797,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06990543007850647,
      "clippedFraction": 0,
      "rawPeak": 0.5779622197151184,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4446.612874991843,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 77591.31887499825,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9192508459091187
        },
        {
          "label": "Solo piano music.",
          "score": 0.06293127685785294
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.017393605783581734
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00035274712718091905
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004135358176426962
        },
        {
          "label": "White noise and static.",
          "score": 0.000030230470656533726
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4890.193540995824,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 71392.87725000759,
      "durationMs": 71415.22041600547,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.42020998670419624,
      "realTimeFactor": 2.3797625750002527,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07891032099723816,
      "clippedFraction": 0,
      "rawPeak": 0.9522222876548767,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4287.244749997626,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 76305.41395700129,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.6612475514411926
        },
        {
          "label": "Solo piano music.",
          "score": 0.30500084161758423
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.024959135800600052
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.007923105731606483
        },
        {
          "label": "White noise and static.",
          "score": 0.0006161691271699965
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00025319939595647156
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4890.193540995824,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 71481.00933300157,
      "durationMs": 71506.8544169917,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.41969189131398443,
      "realTimeFactor": 2.382700311100052,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07430427521467209,
      "clippedFraction": 0,
      "rawPeak": 0.9931477308273315,
      "rawClippedFraction": 0,
      "outputGain": 0.9867615557895109,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4228.392167002312,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 76397.04795798752,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.7284253835678101
        },
        {
          "label": "Solo piano music.",
          "score": 0.25813040137290955
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.007404230069369078
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004535607993602753
        },
        {
          "label": "White noise and static.",
          "score": 0.001100210938602686
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0004041931824758649
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4890.193540995824,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 71664.96162500698,
      "durationMs": 71692.61025000014,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.41861461053977195,
      "realTimeFactor": 2.3888320541668993,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07263593375682831,
      "clippedFraction": 0,
      "rawPeak": 0.917676568031311,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4317.312707993551,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 76582.80379099597,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9909489750862122
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.005155495833605528
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0031548908445984125
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00042622009641490877
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002927964087575674
        },
        {
          "label": "White noise and static.",
          "score": 0.00002158993811463006
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2233.466125006089,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33525.04399999452,
      "durationMs": 33553.32458400517,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8948534116765037,
      "realTimeFactor": 1.1175014666664842,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1061965823173523,
      "clippedFraction": 0,
      "rawPeak": 1.0327708721160889,
      "rawClippedFraction": 0.000005902777777777778,
      "outputGain": 0.9489036014271351,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1512.3305419983808,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 18002051072,
      "totalMs": 35786.79070901126,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9916555881500244
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00394744798541069
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0034396054688841105
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0007712999358773232
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016913263243623078
        },
        {
          "label": "White noise and static.",
          "score": 0.000016916643289732747
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2233.466125006089,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 35316.48570900143,
      "durationMs": 35342.296583999996,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8494616436978505,
      "realTimeFactor": 1.1772161903000475,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10459902137517929,
      "clippedFraction": 0,
      "rawPeak": 1.1305245161056519,
      "rawClippedFraction": 0.000040625,
      "outputGain": 0.8668542663504833,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1542.9354579973733,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 37575.762709006085,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9862956404685974
        },
        {
          "label": "Solo piano music.",
          "score": 0.01336595043540001
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0002147862978745252
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0000848553390824236
        },
        {
          "label": "White noise and static.",
          "score": 0.000021268122509354725
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00001751048694131896
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2233.466125006089,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30627.334041986614,
      "durationMs": 30656.651792000048,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9795171841882611,
      "realTimeFactor": 1.0209111347328872,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06990541517734528,
      "clippedFraction": 0,
      "rawPeak": 0.5779266953468323,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1335.0453749881126,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 32890.11791700614,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9193499088287354
        },
        {
          "label": "Solo piano music.",
          "score": 0.0628088191151619
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.017416654154658318
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0003523124032653868
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004198325768811628
        },
        {
          "label": "White noise and static.",
          "score": 0.00003030549305549357
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2233.466125006089,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33506.61729100102,
      "durationMs": 33530.461000002106,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8953455294950707,
      "realTimeFactor": 1.1168872430333674,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.0789121687412262,
      "clippedFraction": 0,
      "rawPeak": 0.9521446228027344,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1671.8885409936775,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 35763.927125008195,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.6611555218696594
        },
        {
          "label": "Solo piano music.",
          "score": 0.3049392104148865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.025078095495700836
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.007959401234984398
        },
        {
          "label": "White noise and static.",
          "score": 0.0006146704545244575
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002531185164116323
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2233.466125006089,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33312.114541986375,
      "durationMs": 33336.77104199887,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9005732722907215,
      "realTimeFactor": 1.1104038180662126,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07430242002010345,
      "clippedFraction": 0,
      "rawPeak": 0.9931740760803223,
      "rawClippedFraction": 0,
      "outputGain": 0.986735380637083,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1526.1549169954378,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 35570.23716700496,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.7289741039276123
        },
        {
          "label": "Solo piano music.",
          "score": 0.2575244605541229
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.007418065797537565
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004565398208796978
        },
        {
          "label": "White noise and static.",
          "score": 0.0011092673521488905
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00040873762918636203
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2233.466125006089,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32244.135125001776,
      "durationMs": 32270.99229200394,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9304017578296992,
      "realTimeFactor": 1.0748045041667258,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07263577729463577,
      "clippedFraction": 0,
      "rawPeak": 0.9176568388938904,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1464.3199169950094,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 34504.45841701003,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "e432212fec32b8965a14ffa57ae653438d6abd14",
      "fallback": false,
      "grade": "interactive"
    }
  }
]

ACE-Step 1.5 SFT

52 configurations tested
Passed52
Failed0
TestedNov 7, 2026
Runtimeace-step-1.5 c410d249e71ea9385a7b586865e65b1473e1098d

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedFastmacos · aarch6425437 ms / 2438 ms0.848 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6425982 ms / 2438 ms0.866 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6426113 ms / 2438 ms0.870 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6427854 ms / 2438 ms0.928 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431395 ms / 2438 ms1.047 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6437454 ms / 2438 ms1.248 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6443787 ms / 2438 ms1.460 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433963 ms / 2438 ms1.132 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431838 ms / 2438 ms1.061 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431739 ms / 2438 ms1.058 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6434017 ms / 2438 ms1.134 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432111 ms / 2438 ms1.070 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6467027 ms / 2438 ms1.117 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433924 ms / 2438 ms1.131 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433920 ms / 2438 ms1.131 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6434166 ms / 2438 ms1.139 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6434261 ms / 2438 ms1.142 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433508 ms / 2438 ms1.117 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6433148 ms / 2438 ms1.105 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432662 ms / 2438 ms1.089 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430944 ms / 2438 ms1.031 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432407 ms / 2438 ms1.080 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6434258 ms / 2438 ms1.142 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6432958 ms / 6366 ms1.099 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430728 ms / 2468 ms1.024 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431393 ms / 2557 ms1.046 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430714 ms / 2557 ms1.024 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6430750 ms / 2557 ms1.025 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431204 ms / 2557 ms1.040 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431078 ms / 2557 ms1.036 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6431973 ms / 2557 ms1.066 RTF
FP32cpuPassedNot measuredmacos · aarch6465871 ms / 5095 ms2.196 RTF
FP32cpuPassedNot measuredmacos · aarch6468135 ms / 5095 ms2.271 RTF
FP32cpuPassedNot measuredmacos · aarch6469096 ms / 5095 ms2.303 RTF
FP32cpuPassedNot measuredmacos · aarch6468309 ms / 5095 ms2.277 RTF
FP32cpuPassedNot measuredmacos · aarch6469817 ms / 5095 ms2.327 RTF
FP32cpuPassedNot measuredmacos · aarch6470964 ms / 5095 ms2.365 RTF
FP32cpuPassedNot measuredmacos · aarch6474171 ms / 5095 ms2.472 RTF
FP32cpuPassedNot measuredmacos · aarch6470298 ms / 5095 ms2.343 RTF
FP32cpuPassedNot measuredmacos · aarch6469327 ms / 5095 ms2.311 RTF
FP32cpuPassedNot measuredmacos · aarch6468600 ms / 5095 ms2.287 RTF
FP32cpuPassedNot measuredmacos · aarch6467709 ms / 5095 ms2.257 RTF
FP32cpuPassedNot measuredmacos · aarch6466811 ms / 5095 ms2.227 RTF
FP32cpuPassedNot measuredmacos · aarch64126381 ms / 5095 ms2.106 RTF
FP32cpuPassedNot measuredmacos · aarch6466974 ms / 5139 ms2.232 RTF
FP32cpuPassedNot measuredmacos · aarch6468635 ms / 1255 ms2.288 RTF
FP32cpuPassedNot measuredmacos · aarch6473748 ms / 5022 ms2.458 RTF
FP32cpuPassedNot measuredmacos · aarch6472755 ms / 5022 ms2.425 RTF
FP32cpuPassedNot measuredmacos · aarch6472483 ms / 5022 ms2.416 RTF
FP32cpuPassedNot measuredmacos · aarch6470944 ms / 5022 ms2.365 RTF
FP32cpuPassedNot measuredmacos · aarch6470586 ms / 5022 ms2.353 RTF
FP32cpuPassedNot measuredmacos · aarch6470792 ms / 5022 ms2.360 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-07T20:26:10.645276+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Torch2.10.0/Transformers4.57.6/Diffusers0.37.0/exact reviewed ACEc410d249e71ea9385a7b586865e65b1473e1098d", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 c410d249e71ea9385a7b586865e65b1473e1098d
Artifact revision
c410d249e71ea9385a7b586865e65b1473e1098d
Fixture
{
  "file": "model-samples/desktop/ace-step-variants/sft/piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 65871.03295899578,
      "durationMs": 65897.26966699527,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.45543539629437374,
      "realTimeFactor": 2.195701098633193,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 70992.15683399234,
      "postInferenceRssBytes": 14678704128,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 68134.69345899648,
      "durationMs": 68158.82758400403,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.440304321880512,
      "realTimeFactor": 2.271156448633216,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 73253.7147510011,
      "postInferenceRssBytes": 14809300992,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9168417453765869
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.06704521179199219
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.014404719695448875
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0011866034474223852
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000411536602769047
        },
        {
          "label": "White noise and static.",
          "score": 0.00011010244634235278
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69095.8591249946,
      "durationMs": 69123.23533299787,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4341794194313427,
      "realTimeFactor": 2.3031953041664868,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1309482902288437,
      "clippedFraction": 0,
      "rawPeak": 0.9277427196502686,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74218.12249999493,
      "postInferenceRssBytes": 14812217344,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9937844276428223
        },
        {
          "label": "Solo piano music.",
          "score": 0.005519511178135872
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0006109211826696992
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0000680976954754442
        },
        {
          "label": "White noise and static.",
          "score": 0.000010248557373415679
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000006790971838199766
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 68309.40129199007,
      "durationMs": 68336.89716699882,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4391782014274188,
      "realTimeFactor": 2.276980043066336,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.13579370081424713,
      "clippedFraction": 0,
      "rawPeak": 0.9445314407348633,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 73431.78433399589,
      "postInferenceRssBytes": 14815133696,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9511361718177795
        },
        {
          "label": "Solo piano music.",
          "score": 0.02786535769701004
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.020387178286910057
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005249154637567699
        },
        {
          "label": "White noise and static.",
          "score": 0.0000701618118910119
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00001615758810658008
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69817.09783300175,
      "durationMs": 69843.0732919951,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.42969417135839955,
      "realTimeFactor": 2.327236594433392,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1335451751947403,
      "clippedFraction": 0,
      "rawPeak": 1.3016999959945679,
      "rawClippedFraction": 0.000044444444444444447,
      "outputGain": 0.7528616447841563,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74937.96045899217,
      "postInferenceRssBytes": 14825160704,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9663394093513489
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.023365873843431473
        },
        {
          "label": "Solo piano music.",
          "score": 0.007997311651706696
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0017144593875855207
        },
        {
          "label": "White noise and static.",
          "score": 0.0005035583744756877
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00007935748726595193
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70964.03333400667,
      "durationMs": 70989.99233401264,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.42274936458020773,
      "realTimeFactor": 2.365467777800222,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11290287971496582,
      "clippedFraction": 0,
      "rawPeak": 1.6965795755386353,
      "rawClippedFraction": 0.0008857638888888888,
      "outputGain": 0.5776327937278548,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 76084.87950100971,
      "postInferenceRssBytes": 14828191744,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.999769389629364
        },
        {
          "label": "White noise and static.",
          "score": 0.00008508065366186202
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00006298770313151181
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00004162383629591204
        },
        {
          "label": "Solo piano music.",
          "score": 0.000030063611120567657
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000010846447366930079
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 74171.4582919958,
      "durationMs": 74198.13024999166,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4044682508721477,
      "realTimeFactor": 2.472381943066527,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10370178520679474,
      "clippedFraction": 0,
      "rawPeak": 1.382080078125,
      "rawClippedFraction": 0.000010416666666666666,
      "outputGain": 0.7090761349584879,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 79293.01741698873,
      "postInferenceRssBytes": 14862303232,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70298.29883300408,
      "durationMs": 70322.97012499475,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4267528588602974,
      "realTimeFactor": 2.343276627766803,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75417.85729199182,
      "postInferenceRssBytes": 14871494656,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69327.24912499543,
      "durationMs": 69354.68491600477,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4327302810747718,
      "realTimeFactor": 2.3109083041665146,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 74449.57208300184,
      "postInferenceRssBytes": 14877261824,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 68599.82395800762,
      "durationMs": 68625.8140830032,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.43731890650862404,
      "realTimeFactor": 2.286660798600254,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 73720.70125000027,
      "postInferenceRssBytes": 14880407552,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 67708.95762500004,
      "durationMs": 67731.25137499301,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4430728378090281,
      "realTimeFactor": 2.256965254166668,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 72826.13854199008,
      "postInferenceRssBytes": 14891286528,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 66811.16208400636,
      "durationMs": 66838.90695900482,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.44902676535215624,
      "realTimeFactor": 2.227038736133545,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 71933.79412600189,
      "postInferenceRssBytes": 14894694400,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.8030894994735718
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.18385469913482666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.009938768111169338
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0021512957755476236
        },
        {
          "label": "White noise and static.",
          "score": 0.0006145151564851403
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003511632967274636
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5094.887166997069,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 126380.67212500027,
      "durationMs": 126432.78870799986,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.47475613945663586,
      "realTimeFactor": 2.1063445354166714,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.08367056399583817,
      "clippedFraction": 0,
      "rawPeak": 0.7591762542724609,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17215963136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 131527.67587499693,
      "postInferenceRssBytes": 14994980864,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5139.370125005371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 66974.34875000909,
      "durationMs": 67003.48850000591,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.44793268706470146,
      "realTimeFactor": 2.2324782916669696,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17216077824,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 72142.85862501129,
      "postInferenceRssBytes": 14678376448,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9949230551719666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0046165515668690205
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0003071743994951248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010697953985072672
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004124082624912262
        },
        {
          "label": "White noise and static.",
          "score": 0.00000502267084812047
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1254.74795899936,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 68634.95587500802,
      "durationMs": 68663.29295901232,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.43709505772296814,
      "realTimeFactor": 2.2878318625002674,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10806342959403992,
      "clippedFraction": 0,
      "rawPeak": 0.805483877658844,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 17214603264,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 69918.04091801168,
      "postInferenceRssBytes": 14700953600,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 26113.039584000944,
      "durationMs": 26141.91354200011,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1488513201803032,
      "realTimeFactor": 0.8704346528000315,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14004912128,
      "totalMs": 28580.012334001367,
      "postInferenceRssBytes": 2667511808,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25437.159332999727,
      "durationMs": 25464.54362499935,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.179376973948537,
      "realTimeFactor": 0.8479053110999909,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14009106432,
      "totalMs": 27902.642417000607,
      "postInferenceRssBytes": 2678587392,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9174346923828125
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.06650639325380325
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.01434504333883524
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0011906439904123545
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0004141099634580314
        },
        {
          "label": "White noise and static.",
          "score": 0.00010915115854004398
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25982.325249991845,
      "durationMs": 26009.237958001904,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1546310698273403,
      "realTimeFactor": 0.8660775083330615,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.13093340396881104,
      "clippedFraction": 0,
      "rawPeak": 0.9279181361198425,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14009106432,
      "totalMs": 28447.33675000316,
      "postInferenceRssBytes": 2691170304,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9937863349914551
        },
        {
          "label": "Solo piano music.",
          "score": 0.005522475112229586
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0006074464763514698
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00006685480911983177
        },
        {
          "label": "White noise and static.",
          "score": 0.000010197703886660747
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000006766063961549662
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 27854.000708000967,
      "durationMs": 27881.51600000856,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0770445622693836,
      "realTimeFactor": 0.9284666902666989,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.13579340279102325,
      "clippedFraction": 0,
      "rawPeak": 0.9445355534553528,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14019493888,
      "totalMs": 30319.614792009816,
      "postInferenceRssBytes": 2712567808,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9513677358627319
        },
        {
          "label": "Solo piano music.",
          "score": 0.02779480628669262
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.02023148164153099
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005200193845666945
        },
        {
          "label": "White noise and static.",
          "score": 0.0000697817376931198
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000016215124560403638
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31395.16650000587,
      "durationMs": 31417.549667006824,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9555611052419292,
      "realTimeFactor": 1.0465055500001956,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.13355828821659088,
      "clippedFraction": 0,
      "rawPeak": 1.3015730381011963,
      "rawClippedFraction": 0.000044444444444444447,
      "outputGain": 0.7529350803314703,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14019493888,
      "totalMs": 33855.64845900808,
      "postInferenceRssBytes": 2717204480,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9644838571548462
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.024782191962003708
        },
        {
          "label": "Solo piano music.",
          "score": 0.00837866123765707
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0017605037428438663
        },
        {
          "label": "White noise and static.",
          "score": 0.0005124426097609103
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00008241818431997672
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 37454.09691600071,
      "durationMs": 37480.79645799589,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.800980465962957,
      "realTimeFactor": 1.2484698972000237,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11290293186903,
      "clippedFraction": 0,
      "rawPeak": 1.6965821981430054,
      "rawClippedFraction": 0.0008857638888888888,
      "outputGain": 0.5776319008136825,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14035566592,
      "totalMs": 39918.89524999715,
      "postInferenceRssBytes": 2739453952,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9997674822807312
        },
        {
          "label": "White noise and static.",
          "score": 0.000086515887232963
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00006314642814686522
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000042038242099806666
        },
        {
          "label": "Solo piano music.",
          "score": 0.00002977610711241141
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000011035299394279718
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 43786.92520799814,
      "durationMs": 43809.78650000179,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.6851360276496461,
      "realTimeFactor": 1.4595641735999378,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10369989275932312,
      "clippedFraction": 0,
      "rawPeak": 1.3820874691009521,
      "rawClippedFraction": 0.000010416666666666666,
      "outputGain": 0.7090723430388165,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14035566592,
      "totalMs": 46247.88529200305,
      "postInferenceRssBytes": 2744270848,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33963.39970899862,
      "durationMs": 33989.65041700285,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8833037992969675,
      "realTimeFactor": 1.1321133236332874,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 36427.749209004105,
      "postInferenceRssBytes": 2749136896,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31838.440459003323,
      "durationMs": 31867.0610840054,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9422572075610743,
      "realTimeFactor": 1.061281348633444,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 34305.159876006655,
      "postInferenceRssBytes": 2753970176,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31739.44379200111,
      "durationMs": 31766.193582996493,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9451961476262706,
      "realTimeFactor": 1.0579814597333703,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 34204.29237499775,
      "postInferenceRssBytes": 2761670656,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 34016.605249998975,
      "durationMs": 34056.12645900692,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8819222194431322,
      "realTimeFactor": 1.1338868416666326,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 36494.225251008174,
      "postInferenceRssBytes": 2762309632,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32110.69666700496,
      "durationMs": 32138.918666998507,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9342681135543912,
      "realTimeFactor": 1.070356555566832,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14073315328,
      "totalMs": 34577.01745899976,
      "postInferenceRssBytes": 2762850304,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.803131103515625
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.18379409611225128
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.009971600957214832
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0021398563403636217
        },
        {
          "label": "White noise and static.",
          "score": 0.0006141901831142604
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003491021634545177
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 67026.83245799562,
      "durationMs": 67084.99375000247,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.8951638888440805,
      "realTimeFactor": 1.1171138742999271,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08367033302783966,
      "clippedFraction": 0,
      "rawPeak": 0.7591606378555298,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 69523.09254200372,
      "postInferenceRssBytes": 1290158080,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33924.45312500058,
      "durationMs": 33958.61070799583,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8843178662146668,
      "realTimeFactor": 1.130815104166686,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 36396.70949999709,
      "postInferenceRssBytes": 1351745536,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33920.042375000776,
      "durationMs": 33947.17254200077,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8844328573749111,
      "realTimeFactor": 1.1306680791666925,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 36385.27133400203,
      "postInferenceRssBytes": 1360953344,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 34166.374749998795,
      "durationMs": 34194.332083003246,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8780562825150496,
      "realTimeFactor": 1.1388791583332931,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 36632.4308750045,
      "postInferenceRssBytes": 1362477056,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 34261.46437499847,
      "durationMs": 34285.20658399793,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.875619315965135,
      "realTimeFactor": 1.1420488124999488,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 36723.30537599919,
      "postInferenceRssBytes": 1373896704,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33508.076167010586,
      "durationMs": 33536.636458011344,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8953065479042822,
      "realTimeFactor": 1.1169358722336862,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 35974.7352500126,
      "postInferenceRssBytes": 1383022592,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 33147.93987500889,
      "durationMs": 33170.68258400832,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9050336193778907,
      "realTimeFactor": 1.104931329166963,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15274393600,
      "totalMs": 35608.781376009574,
      "postInferenceRssBytes": 1392214016,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32661.8495839939,
      "durationMs": 32689.579499987303,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9185027909350746,
      "realTimeFactor": 1.0887283194664632,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 35127.67829198856,
      "postInferenceRssBytes": 1401192448,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30944.19079099316,
      "durationMs": 30972.525499993935,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9694873006254866,
      "realTimeFactor": 1.0314730263664387,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 33410.62429199519,
      "postInferenceRssBytes": 1410170880,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32407.405542006018,
      "durationMs": 32430.473625005106,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9257143389992891,
      "realTimeFactor": 1.0802468514002006,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 34868.57241700636,
      "postInferenceRssBytes": 1419100160,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2438.098792001256,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 34257.825207998394,
      "durationMs": 34283.77345799527,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.8757123319373965,
      "realTimeFactor": 1.1419275069332797,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15005188096,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15312142336,
      "totalMs": 36721.87224999652,
      "postInferenceRssBytes": 1428111360,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6365.757999999914,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 32958.07070899173,
      "durationMs": 32991.60245899111,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9102474554681778,
      "realTimeFactor": 1.098602356966391,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15017230336,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14004912128,
      "totalMs": 39357.360458991025,
      "postInferenceRssBytes": 2684485632,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9948381781578064
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004717690404504538
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00029463652754202485
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010461978672537953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000040073067793855444
        },
        {
          "label": "White noise and static.",
          "score": 0.000004909230028715683
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2468.0608750059037,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30728.276458001346,
      "durationMs": 30761.6352920013,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9762994693504292,
      "realTimeFactor": 1.0242758819333782,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10806325823068619,
      "clippedFraction": 0,
      "rawPeak": 0.8054794669151306,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 15004581888,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 14004912128,
      "totalMs": 33229.696167007205,
      "postInferenceRssBytes": 2665889792,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9800806045532227
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.01760890707373619
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0018174260621890426
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00034625717671588063
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00013571276213042438
        },
        {
          "label": "White noise and static.",
          "score": 0.000010996649507433176
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5021.64691699727,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 73748.23679099791,
      "durationMs": 73777.17062500597,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4067893865045185,
      "realTimeFactor": 2.4582745596999303,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10434228181838989,
      "clippedFraction": 0,
      "rawPeak": 1.0472356081008911,
      "rawClippedFraction": 0.000020486111111111113,
      "outputGain": 0.9357970569556745,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4846.297666008468,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 78798.81754200323,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 15083651072,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9618105888366699
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.033093784004449844
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002377222292125225
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0023162728175520897
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00024928036145865917
        },
        {
          "label": "White noise and static.",
          "score": 0.00015273207100108266
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5021.64691699727,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 72754.54929099942,
      "durationMs": 72783.16983301193,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4123453487424923,
      "realTimeFactor": 2.425151643033314,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.12284808605909348,
      "clippedFraction": 0,
      "rawPeak": 1.0816426277160645,
      "rawClippedFraction": 0.00006458333333333334,
      "outputGain": 0.906029380581378,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4211.721041996498,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 77804.8167500092,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 16970825728,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9938083291053772
        },
        {
          "label": "Solo piano music.",
          "score": 0.005587448365986347
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0005281433695927262
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00006681524246232584
        },
        {
          "label": "White noise and static.",
          "score": 0.000006819672307756264
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000002384419758527656
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5021.64691699727,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 72483.24145800143,
      "durationMs": 72509.88191599026,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.41388877479193226,
      "realTimeFactor": 2.4161080486000475,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.09388642013072968,
      "clippedFraction": 0,
      "rawPeak": 0.8747938275337219,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4514.503832993796,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 77531.52883298753,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 16199745536,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.862688422203064
        },
        {
          "label": "Solo piano music.",
          "score": 0.11088230460882187
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.025812065228819847
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0003598149633035064
        },
        {
          "label": "White noise and static.",
          "score": 0.00016742016305215657
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00009002436854643747
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5021.64691699727,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70944.19399999606,
      "durationMs": 70972.00995900494,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4228675851895881,
      "realTimeFactor": 2.3648064666665354,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10040242969989777,
      "clippedFraction": 0,
      "rawPeak": 0.9137781858444214,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4247.587416000897,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75993.6568760022,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 15931899904,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.8784697651863098
        },
        {
          "label": "Solo piano music.",
          "score": 0.11286214739084244
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004033456556499004
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.003694212529808283
        },
        {
          "label": "White noise and static.",
          "score": 0.0006787533638998866
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002616934652905911
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5021.64691699727,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70585.56962500734,
      "durationMs": 70613.45183300728,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4250160501555474,
      "realTimeFactor": 2.352852320833578,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07115954905748367,
      "clippedFraction": 0,
      "rawPeak": 1.2972151041030884,
      "rawClippedFraction": 0.000013541666666666666,
      "outputGain": 0.7554645308247354,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4242.52883400186,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75635.09875000454,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 15014969344,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.760319709777832
        },
        {
          "label": "Solo piano music.",
          "score": 0.2090287208557129
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.015240943990647793
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.013674873858690262
        },
        {
          "label": "White noise and static.",
          "score": 0.0012925489572808146
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00044324653572402894
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 5021.64691699727,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 70791.86512499291,
      "durationMs": 70814.32162500278,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4237775053253762,
      "realTimeFactor": 2.359728837499764,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.09822630882263184,
      "clippedFraction": 0,
      "rawPeak": 1.0927493572235107,
      "rawClippedFraction": 0.000004513888888888889,
      "outputGain": 0.8968204771952577,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4434.811583996634,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 75835.96854200005,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 14984904704,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9799964427947998
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.01764586567878723
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0018590440740808845
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0003496142744552344
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0001378787128487602
        },
        {
          "label": "White noise and static.",
          "score": 0.00001110882203647634
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2557.0357919932576,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31393.044000011287,
      "durationMs": 31427.27987500257,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.955625711224092,
      "realTimeFactor": 1.0464348000003763,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10434377938508987,
      "clippedFraction": 0,
      "rawPeak": 1.0471925735473633,
      "rawClippedFraction": 0.000020486111111111113,
      "outputGain": 0.9358355136918622,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1747.9064170038328,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 18027216896,
      "totalMs": 33984.315666995826,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 2803499008,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9618737697601318
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.03301496058702469
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0023652163799852133
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.002340685809031129
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002509454498067498
        },
        {
          "label": "White noise and static.",
          "score": 0.00015444857126567513
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2557.0357919932576,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30713.786959007848,
      "durationMs": 30746.547750008176,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9767600472074478,
      "realTimeFactor": 1.023792898633595,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.12285532057285309,
      "clippedFraction": 0,
      "rawPeak": 1.0815999507904053,
      "rawClippedFraction": 0.00006458333333333334,
      "outputGain": 0.9060651299806748,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1436.7606250016252,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 33303.583542001434,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 2862137344,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9939678907394409
        },
        {
          "label": "Solo piano music.",
          "score": 0.005421042442321777
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0005355653120204806
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00006651003786828369
        },
        {
          "label": "White noise and static.",
          "score": 0.000006683007541141706
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000002358165374971577
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2557.0357919932576,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 30750.29891698796,
      "durationMs": 30775.031458993908,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9756002724066706,
      "realTimeFactor": 1.0250099638995986,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09388714283704758,
      "clippedFraction": 0,
      "rawPeak": 0.8748162984848022,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1323.4075000073062,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 33332.067250987166,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 2910093312,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.8627902269363403
        },
        {
          "label": "Solo piano music.",
          "score": 0.11066517978906631
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.02592652104794979
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00036100068246014416
        },
        {
          "label": "White noise and static.",
          "score": 0.00016692413191776723
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000902395840967074
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2557.0357919932576,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31203.640499996254,
      "durationMs": 31231.54904200055,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9614262797318025,
      "realTimeFactor": 1.0401213499998752,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10040178149938583,
      "clippedFraction": 0,
      "rawPeak": 0.9137697815895081,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1370.4206669935957,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 33788.58483399381,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 2957049856,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.8829342126846313
        },
        {
          "label": "Solo piano music.",
          "score": 0.10852303355932236
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.004038706421852112
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.003596130758523941
        },
        {
          "label": "White noise and static.",
          "score": 0.0006505702040158212
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002573319070506841
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2557.0357919932576,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31077.649041006225,
      "durationMs": 31105.97754099581,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9653239844627792,
      "realTimeFactor": 1.0359216347002076,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07118381559848785,
      "clippedFraction": 0,
      "rawPeak": 1.2967642545700073,
      "rawClippedFraction": 0.000013194444444444444,
      "outputGain": 0.7557271852199204,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1332.0760420028819,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 33663.01333298907,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 3007152128,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.7606394290924072
        },
        {
          "label": "Solo piano music.",
          "score": 0.20868538320064545
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.015236355364322662
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.013699580915272236
        },
        {
          "label": "White noise and static.",
          "score": 0.001296299509704113
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0004429699038155377
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2557.0357919932576,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 31972.84674999537,
      "durationMs": 32000.895457997103,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.9382961809618765,
      "realTimeFactor": 1.065761558333179,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09822633862495422,
      "clippedFraction": 0,
      "rawPeak": 1.0927391052246094,
      "rawClippedFraction": 0.000004513888888888889,
      "outputGain": 0.8968288910998237,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1564.2478749941802,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "diffusion_guidance_sale": 7,
        "infer_steps": 50
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 19503611904,
      "totalMs": 34557.93124999036,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "c410d249e71ea9385a7b586865e65b1473e1098d",
      "postInferenceRssBytes": 3058122752,
      "fallback": false,
      "grade": "interactive"
    }
  }
]

ACE-Step 1.5 Turbo Shift1

40 configurations tested
Passed40
Failed0
TestedNov 7, 2026
Runtimeace-step-1.5 5b86586cc1faecd5214281b05440c8903b6da20f

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedFastmacos · aarch645880 ms / 4965 ms0.196 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645896 ms / 4965 ms0.197 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645919 ms / 4965 ms0.197 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645924 ms / 4965 ms0.197 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645927 ms / 4965 ms0.198 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645928 ms / 4965 ms0.198 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645957 ms / 4965 ms0.199 RTF
FP32mpsAcceleratedPassedFastmacos · aarch645968 ms / 4965 ms0.199 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646032 ms / 4965 ms0.201 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646085 ms / 4965 ms0.203 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646204 ms / 4965 ms0.207 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412518 ms / 4965 ms0.209 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646341 ms / 4965 ms0.211 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646491 ms / 4965 ms0.216 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646517 ms / 4965 ms0.217 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646557 ms / 4965 ms0.219 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646708 ms / 4965 ms0.224 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646932 ms / 4965 ms0.231 RTF
FP32mpsAcceleratedPassedFastmacos · aarch646955 ms / 4965 ms0.232 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647260 ms / 4965 ms0.242 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647468 ms / 4965 ms0.249 RTF
FP32mpsAcceleratedPassedFastmacos · aarch647862 ms / 4965 ms0.262 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648355 ms / 6373 ms0.279 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648662 ms / 3331 ms0.289 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649521 ms / 4965 ms0.317 RTF
FP32cpuPassedFastmacos · aarch6429357 ms / 6911 ms0.489 RTF
FP32cpuPassedFastmacos · aarch6415172 ms / 6911 ms0.506 RTF
FP32cpuPassedFastmacos · aarch6415234 ms / 6911 ms0.508 RTF
FP32cpuPassedFastmacos · aarch6415262 ms / 6911 ms0.509 RTF
FP32cpuPassedFastmacos · aarch6415288 ms / 6911 ms0.510 RTF
FP32cpuPassedFastmacos · aarch6415333 ms / 6911 ms0.511 RTF
FP32cpuPassedFastmacos · aarch6415388 ms / 6911 ms0.513 RTF
FP32cpuPassedFastmacos · aarch6415470 ms / 6911 ms0.516 RTF
FP32cpuPassedFastmacos · aarch6415576 ms / 4948 ms0.519 RTF
FP32cpuPassedFastmacos · aarch6415606 ms / 2589 ms0.520 RTF
FP32cpuPassedFastmacos · aarch6415637 ms / 6911 ms0.521 RTF
FP32cpuPassedFastmacos · aarch6415677 ms / 6911 ms0.523 RTF
FP32cpuPassedFastmacos · aarch6416023 ms / 6911 ms0.534 RTF
FP32cpuPassedFastmacos · aarch6416529 ms / 6911 ms0.551 RTF
FP32cpuPassedFastmacos · aarch6416768 ms / 6911 ms0.559 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-07T17:52:09.849064+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Torch2.10.0/Transformers4.57.6/Diffusers0.37.0/exact reviewed ACE5b86586cc1faecd5214281b05440c8903b6da20f", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 5b86586cc1faecd5214281b05440c8903b6da20f
Artifact revision
5b86586cc1faecd5214281b05440c8903b6da20f
Fixture
{
  "file": "model-samples/desktop/ace-step-variants/turbo-shift1/piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16529.023124996456,
      "durationMs": 16562.538875004975,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.8149892932651113,
      "realTimeFactor": 0.5509674374998819,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 23473.254458018346,
      "postInferenceRssBytes": 13730627584,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16768.35383300204,
      "durationMs": 16791.823999999906,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.7890843847150082,
      "realTimeFactor": 0.5589451277667347,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 23702.539583013277,
      "postInferenceRssBytes": 13838794752,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9963921904563904
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0025234841741621494
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000515855266712606
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00045511932694353163
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00008645220805192366
        },
        {
          "label": "White noise and static.",
          "score": 0.000026871930458582938
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 16023.085209002602,
      "durationMs": 16050.55708400323,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.8722985997193875,
      "realTimeFactor": 0.5341028403000867,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.09860066324472427,
      "clippedFraction": 0,
      "rawPeak": 0.9059590101242065,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22961.2726670166,
      "postInferenceRssBytes": 13854146560,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9686466455459595
        },
        {
          "label": "Solo piano music.",
          "score": 0.031080015003681183
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018761484534479678
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00007940817886264995
        },
        {
          "label": "White noise and static.",
          "score": 0.000003945523076254176
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000024681621653144248
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15637.462958999095,
      "durationMs": 15665.642874999321,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.918469772152874,
      "realTimeFactor": 0.5212487652999698,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.057052016258239746,
      "clippedFraction": 0,
      "rawPeak": 0.4871351420879364,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22576.35845801269,
      "postInferenceRssBytes": 13857587200,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.8873266577720642
        },
        {
          "label": "Solo piano music.",
          "score": 0.11194594949483871
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0004275331739336252
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0002148095372831449
        },
        {
          "label": "White noise and static.",
          "score": 0.000044291180529398844
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004069522401550785
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15676.729667000473,
      "durationMs": 15704.66524999938,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9136644336701183,
      "realTimeFactor": 0.5225576555666824,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0805157795548439,
      "clippedFraction": 0,
      "rawPeak": 0.6576061248779297,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22615.38083301275,
      "postInferenceRssBytes": 13862256640,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.859381377696991
        },
        {
          "label": "Solo piano music.",
          "score": 0.1341879665851593
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0030938759446144104
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017234167316928506
        },
        {
          "label": "White noise and static.",
          "score": 0.0011625149054452777
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0004508198471739888
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15470.350542003871,
      "durationMs": 15498.27241699677,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9391932922622779,
      "realTimeFactor": 0.5156783514001291,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1097455620765686,
      "clippedFraction": 0,
      "rawPeak": 1.1630524396896362,
      "rawClippedFraction": 0.0000020833333333333334,
      "outputGain": 0.8426103299877997,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22408.98800001014,
      "postInferenceRssBytes": 13866467328,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.7155359983444214
        },
        {
          "label": "Solo piano music.",
          "score": 0.2660316526889801
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.006755632348358631
        },
        {
          "label": "White noise and static.",
          "score": 0.0051324269734323025
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.005118583794683218
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.001425687805749476
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15333.428541009198,
      "durationMs": 15362.501750001684,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9565095907784167,
      "realTimeFactor": 0.5111142847003066,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.13382363319396973,
      "clippedFraction": 0,
      "rawPeak": 1.0709095001220703,
      "rawClippedFraction": 0.0000010416666666666667,
      "outputGain": 0.9151100068570615,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22273.217333015054,
      "postInferenceRssBytes": 13892665344,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15388.361707999138,
      "durationMs": 15415.759874987998,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9495252691133116,
      "realTimeFactor": 0.512945390266638,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22326.47545800137,
      "postInferenceRssBytes": 13896073216,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15288.16637500131,
      "durationMs": 15311.804541997844,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9623020357140395,
      "realTimeFactor": 0.509605545833377,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22222.520125011215,
      "postInferenceRssBytes": 13901889536,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15262.278499998502,
      "durationMs": 15289.645166994887,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9656304921970167,
      "realTimeFactor": 0.5087426166666167,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22200.360750008258,
      "postInferenceRssBytes": 13904986112,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15171.542417010642,
      "durationMs": 15198.618084003101,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9773862917433755,
      "realTimeFactor": 0.5057180805670214,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22109.33366701647,
      "postInferenceRssBytes": 13920059392,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15233.917999998084,
      "durationMs": 15261.100792005891,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9692898438867645,
      "realTimeFactor": 0.5077972666666029,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 22171.81637501926,
      "postInferenceRssBytes": 13923532800,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981339573860168
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008116848766803741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0006021141889505088
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0003910336527042091
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000047397814341820776
        },
        {
          "label": "White noise and static.",
          "score": 0.000013942893929197453
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6910.715583013371,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 29356.85295899748,
      "durationMs": 29414.60229200311,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 2.043815802865573,
      "realTimeFactor": 0.489280882649958,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0797346979379654,
      "clippedFraction": 0,
      "rawPeak": 0.6656327843666077,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 36325.31787501648,
      "postInferenceRssBytes": 14023983104,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4947.555874998216,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15575.804292006069,
      "durationMs": 15604.550917007145,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9260642620809525,
      "realTimeFactor": 0.5191934764002023,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 20552.10679200536,
      "postInferenceRssBytes": 14676656128,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985411167144775
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000839819258544594
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002538138360250741
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021391271729953587
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012450673966668546
        },
        {
          "label": "White noise and static.",
          "score": 0.000026837331461138092
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 2588.7102499982575,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 677
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 15605.646291995072,
      "durationMs": 15633.651541997097,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.9223811329998248,
      "realTimeFactor": 0.520188209733169,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06790578365325928,
      "clippedFraction": 0,
      "rawPeak": 0.5904155969619751,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 18222.361791995354,
      "postInferenceRssBytes": 14676754432,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9521.013833000325,
      "durationMs": 9555.174624998472,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.1509249462508344,
      "realTimeFactor": 0.31736712776667747,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13938409472,
      "totalMs": 14520.071167004062,
      "postInferenceRssBytes": 2645590016,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5923.702000000048,
      "durationMs": 5954.812666997896,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.064400606242474,
      "realTimeFactor": 0.19745673333333494,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13938409472,
      "totalMs": 10919.709209003486,
      "postInferenceRssBytes": 2652766208,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9964295029640198
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0024877593386918306
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005120813730172813
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0004572453035507351
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00008655650890432298
        },
        {
          "label": "White noise and static.",
          "score": 0.000026925730708171614
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5968.136833005701,
      "durationMs": 6007.8522080002585,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.026694400518839,
      "realTimeFactor": 0.19893789443352336,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09860072284936905,
      "clippedFraction": 0,
      "rawPeak": 0.9059561491012573,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13938409472,
      "totalMs": 10972.748750005849,
      "postInferenceRssBytes": 2662055936,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9700660109519958
        },
        {
          "label": "Solo piano music.",
          "score": 0.029677731916308403
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001749657531036064
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00007496659964090213
        },
        {
          "label": "White noise and static.",
          "score": 0.000003824151008302579
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000002524302317397087
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5928.257540988852,
      "durationMs": 5953.484249999747,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.0605088919594925,
      "realTimeFactor": 0.19760858469962841,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.05705062672495842,
      "clippedFraction": 0,
      "rawPeak": 0.48713865876197815,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13944406016,
      "totalMs": 10918.380792005337,
      "postInferenceRssBytes": 2680619008,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.8905532956123352
        },
        {
          "label": "Solo piano music.",
          "score": 0.1087266132235527
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0004226623277645558
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00021331419702619314
        },
        {
          "label": "White noise and static.",
          "score": 0.00004390063986647874
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004008762334706262
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5957.431666000048,
      "durationMs": 5981.650375004392,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.0357270854174425,
      "realTimeFactor": 0.1985810555333349,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08051576465368271,
      "clippedFraction": 0,
      "rawPeak": 0.6576088666915894,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13944406016,
      "totalMs": 10946.546917009982,
      "postInferenceRssBytes": 2685730816,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.8582320213317871
        },
        {
          "label": "Solo piano music.",
          "score": 0.13528203964233398
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0031448451336473227
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017137257382273674
        },
        {
          "label": "White noise and static.",
          "score": 0.0011764338705688715
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0004509301797952503
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6084.910749996197,
      "durationMs": 6108.648499997798,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.930228434331391,
      "realTimeFactor": 0.2028303583332066,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10974493622779846,
      "clippedFraction": 0,
      "rawPeak": 1.1630563735961914,
      "rawClippedFraction": 0.0000020833333333333334,
      "outputGain": 0.8426074799537208,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 11073.545042003389,
      "postInferenceRssBytes": 2706292736,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.7209782600402832
        },
        {
          "label": "Solo piano music.",
          "score": 0.2605055272579193
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0067059057764709
        },
        {
          "label": "White noise and static.",
          "score": 0.005262607242912054
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.005096932407468557
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.001450784271582961
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5880.447415998788,
      "durationMs": 5908.485040999949,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.101652625679424,
      "realTimeFactor": 0.19601491386662626,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.13382144272327423,
      "clippedFraction": 0,
      "rawPeak": 1.070914626121521,
      "rawClippedFraction": 0.0000010416666666666667,
      "outputGain": 0.9151056266261093,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 10873.38158300554,
      "postInferenceRssBytes": 2710257664,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5895.608458988136,
      "durationMs": 5922.461208989262,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.088533305542632,
      "realTimeFactor": 0.1965202819662712,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 10887.357750994852,
      "postInferenceRssBytes": 2713403392,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5927.0549999928335,
      "durationMs": 5950.067166995723,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.061535619297657,
      "realTimeFactor": 0.19756849999976112,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 10914.963709001313,
      "postInferenceRssBytes": 2715713536,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 5918.831083996338,
      "durationMs": 5951.2470420013415,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 5.068568366668827,
      "realTimeFactor": 0.1972943694665446,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 10916.143584006932,
      "postInferenceRssBytes": 2721693696,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6031.648457996198,
      "durationMs": 6061.979707999853,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.973764669628382,
      "realTimeFactor": 0.20105494859987327,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 11026.876250005444,
      "postInferenceRssBytes": 2722299904,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6204.250916998717,
      "durationMs": 6233.041542000137,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.835394377394457,
      "realTimeFactor": 0.20680836389995724,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13951320064,
      "totalMs": 11197.938084005727,
      "postInferenceRssBytes": 2720382976,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981196522712708
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008047747542150319
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0006110970862209797
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000403080863179639
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00004715812610811554
        },
        {
          "label": "White noise and static.",
          "score": 0.000014114716577751096
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12518.444417000865,
      "durationMs": 12580.049124997458,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 4.792927779310669,
      "realTimeFactor": 0.20864074028334775,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07973486185073853,
      "clippedFraction": 0,
      "rawPeak": 0.6656450033187866,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026667520,
      "totalMs": 17544.94566700305,
      "postInferenceRssBytes": 2793897984,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6341.157832997851,
      "durationMs": 6368.391541996971,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.730997207463795,
      "realTimeFactor": 0.21137192776659502,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11333.288084002561,
      "postInferenceRssBytes": 2736324608,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6556.975416999194,
      "durationMs": 6583.584791995236,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.575280230916213,
      "realTimeFactor": 0.21856584723330647,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11548.481334000826,
      "postInferenceRssBytes": 2692448256,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6490.66166600096,
      "durationMs": 6517.215500003658,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.622024925000237,
      "realTimeFactor": 0.21635538886669867,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11482.112042009248,
      "postInferenceRssBytes": 2648637440,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6516.681041990523,
      "durationMs": 6547.155916996417,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.603570407496342,
      "realTimeFactor": 0.2172227013996841,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11512.052459002007,
      "postInferenceRssBytes": 2638544896,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6707.871791993966,
      "durationMs": 6736.168624993297,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.472357392967147,
      "realTimeFactor": 0.22359572639979888,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11701.065166998887,
      "postInferenceRssBytes": 2636644352,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6955.258749992936,
      "durationMs": 6983.056832992588,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.313283096769113,
      "realTimeFactor": 0.23184195833309787,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11947.953374998178,
      "postInferenceRssBytes": 2634711040,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 6932.350916991709,
      "durationMs": 6960.029957990628,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.327536265723041,
      "realTimeFactor": 0.23107836389972364,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 11924.926499996218,
      "postInferenceRssBytes": 2624372736,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7260.316708998289,
      "durationMs": 7286.801709007705,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.132051149066075,
      "realTimeFactor": 0.24201055696660964,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 12251.698251013295,
      "postInferenceRssBytes": 2580529152,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7467.7847499988275,
      "durationMs": 7494.6295839909,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 4.017255585735075,
      "realTimeFactor": 0.24892615833329426,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15037153280,
      "totalMs": 12459.52612599649,
      "postInferenceRssBytes": 2536620032,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4964.89654200559,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 7862.130333000096,
      "durationMs": 7890.859833001741,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.8157596897216988,
      "realTimeFactor": 0.2620710111000032,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 15026667520,
      "totalMs": 12855.756375007331,
      "postInferenceRssBytes": 2492743680,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6372.944665999967,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8355.331958009629,
      "durationMs": 8401.033208006993,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.590521615510591,
      "realTimeFactor": 0.2785110652669876,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13927923712,
      "totalMs": 14773.97787400696,
      "postInferenceRssBytes": 2656845824,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9985432624816895
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0008629329968243837
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00024301154189743102
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00020250266243238002
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012250413419678807
        },
        {
          "label": "White noise and static.",
          "score": 0.000025781491785892285
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 3330.6582080113003,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 677
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8661.986541992519,
      "durationMs": 8691.443999996409,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.463408752087381,
      "realTimeFactor": 0.28873288473308395,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06790587306022644,
      "clippedFraction": 0,
      "rawPeak": 0.5904122591018677,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 1
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 13927923712,
      "totalMs": 12022.10220800771,
      "postInferenceRssBytes": 2661613568,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "5b86586cc1faecd5214281b05440c8903b6da20f",
      "fallback": false,
      "grade": "fast"
    }
  }
]

ACE-Step 1.5 XL Base

40 configurations tested
Passed40
Failed0
TestedNov 7, 2026
Runtimeace-step-1.5 220c1166efbdd9583eafcb12eb160594bbfcb241

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedNot measuredmacos · aarch6469496 ms / 11105 ms2.317 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6478279 ms / 11105 ms2.609 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6493625 ms / 11105 ms3.121 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486394 ms / 11105 ms2.880 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6482774 ms / 11105 ms2.759 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6481314 ms / 11105 ms2.710 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6480963 ms / 11105 ms2.699 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6471540 ms / 11105 ms2.385 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6471648 ms / 11105 ms2.388 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6474586 ms / 11105 ms2.486 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6478203 ms / 11105 ms2.607 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6479161 ms / 11105 ms2.639 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch64169714 ms / 11105 ms2.829 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6478050 ms / 11105 ms2.602 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6478371 ms / 11105 ms2.612 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6479853 ms / 11105 ms2.662 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6479847 ms / 11105 ms2.662 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6480219 ms / 11105 ms2.674 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6480050 ms / 11105 ms2.668 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6480096 ms / 11105 ms2.670 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6479949 ms / 11105 ms2.665 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6480471 ms / 11105 ms2.682 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6482425 ms / 11105 ms2.747 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6478713 ms / 10139 ms2.624 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6475616 ms / 4088 ms2.521 RTF
FP32cpuPassedNot measuredmacos · aarch64166164 ms / 658 ms5.539 RTF
FP32cpuPassedNot measuredmacos · aarch64171305 ms / 658 ms5.710 RTF
FP32cpuPassedNot measuredmacos · aarch64172745 ms / 658 ms5.758 RTF
FP32cpuPassedNot measuredmacos · aarch64170559 ms / 658 ms5.685 RTF
FP32cpuPassedNot measuredmacos · aarch64173939 ms / 658 ms5.798 RTF
FP32cpuPassedNot measuredmacos · aarch64170560 ms / 658 ms5.685 RTF
FP32cpuPassedNot measuredmacos · aarch64171540 ms / 658 ms5.718 RTF
FP32cpuPassedNot measuredmacos · aarch64170495 ms / 658 ms5.683 RTF
FP32cpuPassedNot measuredmacos · aarch64170690 ms / 658 ms5.690 RTF
FP32cpuPassedNot measuredmacos · aarch64170419 ms / 658 ms5.681 RTF
FP32cpuPassedNot measuredmacos · aarch64171064 ms / 658 ms5.702 RTF
FP32cpuPassedNot measuredmacos · aarch64170209 ms / 658 ms5.674 RTF
FP32cpuPassedNot measuredmacos · aarch64305211 ms / 658 ms5.087 RTF
FP32cpuPassedNot measuredmacos · aarch64169320 ms / 1383 ms5.644 RTF
FP32cpuPassedNot measuredmacos · aarch64169766 ms / 545 ms5.659 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-07T23:42:58.913824+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Torch2.10.0/Transformers4.57.6/Diffusers0.37.0/exact reviewed ACE220c1166efbdd9583eafcb12eb160594bbfcb241", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 220c1166efbdd9583eafcb12eb160594bbfcb241
Artifact revision
220c1166efbdd9583eafcb12eb160594bbfcb241
Fixture
{
  "file": "model-samples/desktop/ace-step-variants/xl-base/piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 166164.08783399675,
      "durationMs": 166191.17908399494,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1805444268437291,
      "realTimeFactor": 5.538802927799892,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 24982208512,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 166849.63500099548,
      "postInferenceRssBytes": 24982880256,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 171304.91070800053,
      "durationMs": 171333.20733300934,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17512632811289802,
      "realTimeFactor": 5.710163690266684,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25013501952,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171991.66325000988,
      "postInferenceRssBytes": 25013501952,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.7718042135238647
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.1861058622598648
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.03969145193696022
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.002262822352349758
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00007725971954641864
        },
        {
          "label": "White noise and static.",
          "score": 0.0000583696091780439
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 172744.83966699336,
      "durationMs": 172771.73245798622,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1736665480591612,
      "realTimeFactor": 5.758161322233112,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.08043057471513748,
      "clippedFraction": 0,
      "rawPeak": 0.48235681653022766,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 173430.18837498676,
      "postInferenceRssBytes": 24987500544,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9972026348114014
        },
        {
          "label": "Solo piano music.",
          "score": 0.002677417825907469
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00007924064266262576
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00003596010719775222
        },
        {
          "label": "White noise and static.",
          "score": 0.0000027065752874477766
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000001980707338589127
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 170559.36600000132,
      "durationMs": 170587.07666699775,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1758918358080656,
      "realTimeFactor": 5.685312200000044,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06739506870508194,
      "clippedFraction": 0,
      "rawPeak": 0.714052140712738,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171245.53258399828,
      "postInferenceRssBytes": 24998215680,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9530926942825317
        },
        {
          "label": "Solo piano music.",
          "score": 0.030665630474686623
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.01481052115559578
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0013677474344149232
        },
        {
          "label": "White noise and static.",
          "score": 0.00004336180063546635
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000019963279555668123
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 173938.84658400202,
      "durationMs": 173961.95766699384,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17247441034117586,
      "realTimeFactor": 5.797961552800067,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07701785117387772,
      "clippedFraction": 0,
      "rawPeak": 0.45394566655158997,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 174620.41358399438,
      "postInferenceRssBytes": 24931794944,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9980702996253967
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006549509707838297
        },
        {
          "label": "White noise and static.",
          "score": 0.0006064175977371633
        },
        {
          "label": "Solo piano music.",
          "score": 0.00030787556897848845
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002826769195962697
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00007788880611769855
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 170559.62720900425,
      "durationMs": 170587.024249995,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17589156643288106,
      "realTimeFactor": 5.685320906966808,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0822855681180954,
      "clippedFraction": 0,
      "rawPeak": 1.5696974992752075,
      "rawClippedFraction": 0.00011076388888888888,
      "outputGain": 0.6243241136922912,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171245.48016699555,
      "postInferenceRssBytes": 24985174016,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9990038275718689
        },
        {
          "label": "White noise and static.",
          "score": 0.00037619328941218555
        },
        {
          "label": "Solo piano music.",
          "score": 0.0002644367632456124
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00020033182227052748
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0001386269141221419
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000016522757505299523
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 171539.86454199185,
      "durationMs": 171564.95524999627,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17488646199004193,
      "realTimeFactor": 5.717995484733062,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0970706194639206,
      "clippedFraction": 0,
      "rawPeak": 1.6323891878128052,
      "rawClippedFraction": 0.00016631944444444444,
      "outputGain": 0.6003470295665679,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 172223.4111669968,
      "postInferenceRssBytes": 24839929856,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 170495.46995900164,
      "durationMs": 170521.05966700765,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1759577542278043,
      "realTimeFactor": 5.683182331966721,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171179.5155840082,
      "postInferenceRssBytes": 24990777344,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 170690.14941698697,
      "durationMs": 170716.21462499024,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17575706683993575,
      "realTimeFactor": 5.689671647232899,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171374.67054199078,
      "postInferenceRssBytes": 24984256512,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 170418.83500000404,
      "durationMs": 170447.35329100513,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1760368799610635,
      "realTimeFactor": 5.680627833333468,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171105.80920800567,
      "postInferenceRssBytes": 24966103040,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 171063.5075409955,
      "durationMs": 171090.721665998,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17537346469299117,
      "realTimeFactor": 5.702116918033183,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171749.17758299853,
      "postInferenceRssBytes": 24877350912,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 170209.35141699738,
      "durationMs": 170234.14662499272,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17625353572085906,
      "realTimeFactor": 5.673645047233246,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25037373440,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 170892.60254199326,
      "postInferenceRssBytes": 25023119360,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.962063729763031
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.024821070954203606
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.01222261507064104
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0008196349372155964
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000053318817663239315
        },
        {
          "label": "White noise and static.",
          "score": 0.000019581180822569877
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 658.4559170005377,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 305210.88741700805,
      "durationMs": 305266.50383400556,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.19658538562558653,
      "realTimeFactor": 5.086848123616801,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0752430334687233,
      "clippedFraction": 0,
      "rawPeak": 0.4787023067474365,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25318948864,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 305924.9597510061,
      "postInferenceRssBytes": 25119653888,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1383.1004169915104,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 169320.41624998965,
      "durationMs": 169348.95116699045,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17717886988717957,
      "realTimeFactor": 5.644013874999655,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 24984190976,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 170732.05158398196,
      "postInferenceRssBytes": 24984846336,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9962343573570251
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017025949200615287
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008378431666642427
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008030578610487282
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000292581768007949
        },
        {
          "label": "White noise and static.",
          "score": 0.00012947662617079914
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 545.4279589903308,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 169765.7257910032,
      "durationMs": 169791.58491600538,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17671411505601953,
      "realTimeFactor": 5.6588575263667735,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07252166420221329,
      "clippedFraction": 0,
      "rawPeak": 0.37857067584991455,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 24999919616,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 170337.0128749957,
      "postInferenceRssBytes": 25000591360,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 69495.67062499409,
      "durationMs": 69528.85345899267,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.43168156707031635,
      "realTimeFactor": 2.31652235416647,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24804229120,
      "totalMs": 80633.4370419936,
      "postInferenceRssBytes": 1248657408,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 78278.98683300009,
      "durationMs": 78320.13924999046,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.38324461281035505,
      "realTimeFactor": 2.609299561100003,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24766480384,
      "totalMs": 89424.72283299139,
      "postInferenceRssBytes": 1302888448,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.7716774344444275
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.1859716773033142
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.03994277864694595
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.002271440112963319
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00007756349805276841
        },
        {
          "label": "White noise and static.",
          "score": 0.000059107675042469054
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 93624.71579200064,
      "durationMs": 93652.39462499449,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3204282090067847,
      "realTimeFactor": 3.120823859733355,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08043058216571808,
      "clippedFraction": 0,
      "rawPeak": 0.4823588728904724,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24766480384,
      "totalMs": 104756.97820799542,
      "postInferenceRssBytes": 1310048256,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9971994161605835
        },
        {
          "label": "Solo piano music.",
          "score": 0.0026804571971297264
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00007929398998385295
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00003607781764003448
        },
        {
          "label": "White noise and static.",
          "score": 0.0000027234691515332088
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000019680253444676055
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86393.82816699799,
      "durationMs": 86421.68320900237,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.34724702720674033,
      "realTimeFactor": 2.8797942722332666,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06739504635334015,
      "clippedFraction": 0,
      "rawPeak": 0.7140410542488098,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24785649664,
      "totalMs": 97526.2667920033,
      "postInferenceRssBytes": 1333641216,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9533191323280334
        },
        {
          "label": "Solo piano music.",
          "score": 0.030434563755989075
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.014810176566243172
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.001372328377328813
        },
        {
          "label": "White noise and static.",
          "score": 0.000043809555791085586
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00002003583540499676
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 82773.86924999882,
      "durationMs": 82801.75537499599,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.36243322043327614,
      "realTimeFactor": 2.759128974999961,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07701712846755981,
      "clippedFraction": 0,
      "rawPeak": 0.4538736343383789,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24823398400,
      "totalMs": 93906.33895799692,
      "postInferenceRssBytes": 1340506112,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9981244206428528
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006330548203550279
        },
        {
          "label": "White noise and static.",
          "score": 0.0005971830105409026
        },
        {
          "label": "Solo piano music.",
          "score": 0.0002928205649368465
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002726780658122152
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00007980418740771711
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 81314.19816699054,
      "durationMs": 81341.51620899502,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3689392587797599,
      "realTimeFactor": 2.710473272233018,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08245272934436798,
      "clippedFraction": 0,
      "rawPeak": 1.5664150714874268,
      "rawClippedFraction": 0.00011145833333333334,
      "outputGain": 0.625632386867561,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24895537152,
      "totalMs": 92446.09979199595,
      "postInferenceRssBytes": 1367048192,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9990010857582092
        },
        {
          "label": "White noise and static.",
          "score": 0.000378228141926229
        },
        {
          "label": "Solo piano music.",
          "score": 0.00026304260245524347
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00020186533220112324
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00013944650709163398
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000016384707123506814
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 80962.78433299449,
      "durationMs": 80987.76270799863,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3705406162492142,
      "realTimeFactor": 2.6987594777664827,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.09708469361066818,
      "clippedFraction": 0,
      "rawPeak": 1.63213312625885,
      "rawClippedFraction": 0.00016527777777777778,
      "outputGain": 0.6004412166097876,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24895537152,
      "totalMs": 92092.34629099956,
      "postInferenceRssBytes": 1373634560,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 71540.23912500998,
      "durationMs": 71563.4037500131,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.41934441884626306,
      "realTimeFactor": 2.3846746375003325,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 82667.98733301403,
      "postInferenceRssBytes": 1381367808,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 71647.74720800051,
      "durationMs": 71675.42170800152,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.41871518880987324,
      "realTimeFactor": 2.388258240266684,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 82780.00529100245,
      "postInferenceRssBytes": 1387708416,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 74586.17104199948,
      "durationMs": 74614.0239580127,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4022193334352423,
      "realTimeFactor": 2.4862057013999825,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 85718.60754101363,
      "postInferenceRssBytes": 1396850688,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 78203.46029099892,
      "durationMs": 78228.37229099241,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3836147388922245,
      "realTimeFactor": 2.606782009699964,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 89332.95587399334,
      "postInferenceRssBytes": 1407336448,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 79161.40483399795,
      "durationMs": 79184.65783400461,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.378972556928597,
      "realTimeFactor": 2.6387134944665984,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 90289.24141700554,
      "postInferenceRssBytes": 1417953280,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9621273875236511
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.024642523378133774
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.012333575636148453
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0008231712854467332
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000053671770729124546
        },
        {
          "label": "White noise and static.",
          "score": 0.000019645513020805083
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 169713.54475000408,
      "durationMs": 169774.66708300926,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.3535368970601774,
      "realTimeFactor": 2.8285590791667348,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07523871213197708,
      "clippedFraction": 0,
      "rawPeak": 0.478689581155777,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 180879.2506660102,
      "postInferenceRssBytes": 1718009856,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 78049.60374999791,
      "durationMs": 78078.40758400562,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3843709456372583,
      "realTimeFactor": 2.601653458333264,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 89182.99116700655,
      "postInferenceRssBytes": 1731657728,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 78370.78387499787,
      "durationMs": 78397.91829198657,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.38279571182866157,
      "realTimeFactor": 2.612359462499929,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 89502.5018749875,
      "postInferenceRssBytes": 1742110720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 79853.21008399478,
      "durationMs": 79878.81754199043,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.37568934258803194,
      "realTimeFactor": 2.6617736694664926,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 90983.40112499136,
      "postInferenceRssBytes": 1752580096,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 79847.3025419953,
      "durationMs": 79871.81595899165,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.375717138149052,
      "realTimeFactor": 2.6615767513998434,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 90976.39954199258,
      "postInferenceRssBytes": 1763098624,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 80218.75395800453,
      "durationMs": 80246.46154200309,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.37397738707965167,
      "realTimeFactor": 2.6739584652668174,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 91351.04512500402,
      "postInferenceRssBytes": 1773535232,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 80049.70437500742,
      "durationMs": 80076.7978750082,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3747671554095882,
      "realTimeFactor": 2.668323479166914,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 91181.38145800913,
      "postInferenceRssBytes": 1784070144,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 80096.11283401318,
      "durationMs": 80122.24345900177,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.37455001171118457,
      "realTimeFactor": 2.6698704278004395,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 91226.8270420027,
      "postInferenceRssBytes": 1794572288,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 79948.66379199084,
      "durationMs": 79971.01870799088,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3752407929925323,
      "realTimeFactor": 2.664955459733028,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 91075.60229099181,
      "postInferenceRssBytes": 1805139968,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 80471.24262500438,
      "durationMs": 80498.99258300138,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.37280398588847274,
      "realTimeFactor": 2.682374754166813,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 91603.57616600231,
      "postInferenceRssBytes": 1815658496,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 11104.583583000931,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 82424.97166700196,
      "durationMs": 82453.66833399748,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.363967368059287,
      "realTimeFactor": 2.747499055566732,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4031643648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 93558.25191699842,
      "postInferenceRssBytes": 1826193408,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10138.877000004868,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 78712.95983299206,
      "durationMs": 78744.23729199043,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.38113164672821365,
      "realTimeFactor": 2.6237653277664017,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4043210752,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24804229120,
      "totalMs": 88883.1142919953,
      "postInferenceRssBytes": 1893728256,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9961833357810974
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0017318186582997441
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0008351295255124569
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0008213251130655408
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002987482585012913
        },
        {
          "label": "White noise and static.",
          "score": 0.0001296285045100376
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4088.375082996208,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 75616.4778329985,
      "durationMs": 75645.76633300749,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3967389233105514,
      "realTimeFactor": 2.52054926109995,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07252080738544464,
      "clippedFraction": 0,
      "rawPeak": 0.378690630197525,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4041211904,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24804229120,
      "totalMs": 79734.1414160037,
      "postInferenceRssBytes": 1898889216,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "220c1166efbdd9583eafcb12eb160594bbfcb241",
      "fallback": false,
      "grade": "interactive"
    }
  }
]

ACE-Step 1.5 XL SFT

52 configurations tested
Passed52
Failed0
TestedNov 8, 2026
Runtimeace-step-1.5 d06de46b4622f781cf07f4a013a67d591ca52819

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedNot measuredmacos · aarch6464823 ms / 10333 ms2.161 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6466233 ms / 10333 ms2.208 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6488510 ms / 10333 ms2.950 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6495964 ms / 10333 ms3.199 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6490754 ms / 10333 ms3.025 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6488118 ms / 10333 ms2.937 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6493010 ms / 10333 ms3.100 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485350 ms / 10333 ms2.845 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6484916 ms / 10333 ms2.831 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485314 ms / 10333 ms2.844 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486636 ms / 10333 ms2.888 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485791 ms / 10333 ms2.860 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch64187036 ms / 10333 ms3.117 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6484635 ms / 10333 ms2.821 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6484950 ms / 10333 ms2.832 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486190 ms / 10333 ms2.873 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486155 ms / 10333 ms2.872 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485357 ms / 10333 ms2.845 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485568 ms / 10333 ms2.852 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485310 ms / 10333 ms2.844 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6485441 ms / 10333 ms2.848 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486179 ms / 10333 ms2.873 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486269 ms / 10333 ms2.876 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6482585 ms / 12756 ms2.753 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6481995 ms / 4813 ms2.733 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6483187 ms / 4150 ms2.773 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6484777 ms / 4150 ms2.826 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486222 ms / 4150 ms2.874 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486105 ms / 4150 ms2.870 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6487957 ms / 4150 ms2.932 RTF
FP32mpsAcceleratedPassedNot measuredmacos · aarch6486857 ms / 4150 ms2.895 RTF
FP32cpuPassedNot measuredmacos · aarch64197210 ms / 1453 ms6.574 RTF
FP32cpuPassedNot measuredmacos · aarch64168873 ms / 1453 ms5.629 RTF
FP32cpuPassedNot measuredmacos · aarch64169734 ms / 1453 ms5.658 RTF
FP32cpuPassedNot measuredmacos · aarch64168021 ms / 1453 ms5.601 RTF
FP32cpuPassedNot measuredmacos · aarch64168894 ms / 1453 ms5.630 RTF
FP32cpuPassedNot measuredmacos · aarch64168423 ms / 1453 ms5.614 RTF
FP32cpuPassedNot measuredmacos · aarch64169939 ms / 1453 ms5.665 RTF
FP32cpuPassedNot measuredmacos · aarch64167555 ms / 1453 ms5.585 RTF
FP32cpuPassedNot measuredmacos · aarch64168365 ms / 1453 ms5.612 RTF
FP32cpuPassedNot measuredmacos · aarch64164442 ms / 1453 ms5.481 RTF
FP32cpuPassedNot measuredmacos · aarch64165234 ms / 1453 ms5.508 RTF
FP32cpuPassedNot measuredmacos · aarch64165798 ms / 1453 ms5.527 RTF
FP32cpuPassedNot measuredmacos · aarch64307559 ms / 1453 ms5.126 RTF
FP32cpuPassedNot measuredmacos · aarch64171733 ms / 1402 ms5.724 RTF
FP32cpuPassedNot measuredmacos · aarch64171228 ms / 594 ms5.708 RTF
FP32cpuPassedNot measuredmacos · aarch64204639 ms / 561 ms6.821 RTF
FP32cpuPassedNot measuredmacos · aarch64175145 ms / 561 ms5.838 RTF
FP32cpuPassedNot measuredmacos · aarch64175139 ms / 561 ms5.838 RTF
FP32cpuPassedNot measuredmacos · aarch64174929 ms / 561 ms5.831 RTF
FP32cpuPassedNot measuredmacos · aarch64175071 ms / 561 ms5.836 RTF
FP32cpuPassedNot measuredmacos · aarch64174210 ms / 561 ms5.807 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-08T08:04:08.342018+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Torch2.10.0/Transformers4.57.6/Diffusers0.37.0/exact reviewed ACEd06de46b4622f781cf07f4a013a67d591ca52819", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 d06de46b4622f781cf07f4a013a67d591ca52819
Artifact revision
d06de46b4622f781cf07f4a013a67d591ca52819
Fixture
{
  "file": "model-samples/desktop/ace-step-variants/xl-sft/piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 197210.20304199998,
      "durationMs": 197237.37245899998,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.15212194672103693,
      "realTimeFactor": 6.573673434733332,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 24956862464,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 198690.37462499994,
      "postInferenceRssBytes": 24957534208,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 168872.60879199993,
      "durationMs": 168900.100125,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1776487034493021,
      "realTimeFactor": 5.629086959733331,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 24995168256,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 170353.10229099996,
      "postInferenceRssBytes": 24861261824,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9152688384056091
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.07833848148584366
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.005966509226709604
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00039848449523560703
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000021204798031249084
        },
        {
          "label": "White noise and static.",
          "score": 0.0000064966084210027475
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 169734.43854200002,
      "durationMs": 169760.92870900003,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17674668887290446,
      "realTimeFactor": 5.657814618066667,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.12469033896923065,
      "clippedFraction": 0,
      "rawPeak": 1.1111738681793213,
      "rawClippedFraction": 0.000008680555555555556,
      "outputGain": 0.881950186252803,
      "peakResidentMemoryBytes": 24995168256,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171213.930875,
      "postInferenceRssBytes": 24872452096,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.997431218624115
        },
        {
          "label": "Solo piano music.",
          "score": 0.002438630210235715
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0000798988839960657
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000033625809010118246
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00000990822536550695
        },
        {
          "label": "White noise and static.",
          "score": 0.000006822935119998874
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 168020.810417,
      "durationMs": 168046.1590000001,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17854931139508812,
      "realTimeFactor": 5.600693680566667,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11532054841518402,
      "clippedFraction": 0,
      "rawPeak": 0.7632580995559692,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 24995168256,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 169499.16116600006,
      "postInferenceRssBytes": 24849580032,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9687804579734802
        },
        {
          "label": "Solo piano music.",
          "score": 0.02357564866542816
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.007388773839920759
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00019072751456405967
        },
        {
          "label": "White noise and static.",
          "score": 0.00003680562804220244
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00002757424044830259
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 168893.69079200015,
      "durationMs": 168917.79654200014,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17762652861287928,
      "realTimeFactor": 5.629789693066672,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.1588314175605774,
      "clippedFraction": 0,
      "rawPeak": 1.113883137702942,
      "rawClippedFraction": 0.0001420138888888889,
      "outputGain": 0.8798050413268337,
      "peakResidentMemoryBytes": 24995168256,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 170370.7987080001,
      "postInferenceRssBytes": 24874500096,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9986573457717896
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006478637806139886
        },
        {
          "label": "Solo piano music.",
          "score": 0.0005754158482886851
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00006564849172718823
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000030857929232297465
        },
        {
          "label": "White noise and static.",
          "score": 0.00002281977140228264
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 168422.7290000001,
      "durationMs": 168445.36625000002,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17812322706159203,
      "realTimeFactor": 5.61409096666667,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15270623564720154,
      "clippedFraction": 0,
      "rawPeak": 1.64741051197052,
      "rawClippedFraction": 0.0004163194444444444,
      "outputGain": 0.5948729796726809,
      "peakResidentMemoryBytes": 24995168256,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 169898.36841599998,
      "postInferenceRssBytes": 24890327040,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9988813996315002
        },
        {
          "label": "Solo piano music.",
          "score": 0.0007049881969578564
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00016825823695398867
        },
        {
          "label": "White noise and static.",
          "score": 0.00011725198419298977
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010501022188691422
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00002318220867891796
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 169938.70712500007,
      "durationMs": 169961.39508300007,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17653423700542342,
      "realTimeFactor": 5.664623570833336,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.10898984968662262,
      "clippedFraction": 0,
      "rawPeak": 2.2378909587860107,
      "rawClippedFraction": 0.007372222222222222,
      "outputGain": 0.4379123103172198,
      "peakResidentMemoryBytes": 24995168256,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171414.39724900003,
      "postInferenceRssBytes": 24877678592,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 167555.12037500012,
      "durationMs": 167582.30787499997,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17904555786094686,
      "realTimeFactor": 5.58517067916667,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 25045876736,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 169035.31004099993,
      "postInferenceRssBytes": 25045860352,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 168364.60608400012,
      "durationMs": 168389.35579200007,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17818471885375042,
      "realTimeFactor": 5.612153536133337,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 25073483776,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 169842.35795800004,
      "postInferenceRssBytes": 25020252160,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 164442.166583,
      "durationMs": 164469.49508299987,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1824349594959751,
      "realTimeFactor": 5.481405552766667,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 25073483776,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 165922.49724899983,
      "postInferenceRssBytes": 24977899520,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 165233.60387499997,
      "durationMs": 165259.95758399996,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.18156113100755913,
      "realTimeFactor": 5.507786795833332,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 25073483776,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 166712.95974999992,
      "postInferenceRssBytes": 25032556544,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 165797.95387500006,
      "durationMs": 165825.0806660003,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.18094312564688148,
      "realTimeFactor": 5.526598462500002,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 25073483776,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 167278.08283200028,
      "postInferenceRssBytes": 24968806400,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.8650391101837158
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.09568336606025696
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.03825012221932411
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000857855542562902
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000873026656336151
        },
        {
          "label": "White noise and static.",
          "score": 0.00008217869617510587
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1453.0021659999761,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 307558.74220899976,
      "durationMs": 307615.5166669996,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.1950846838852896,
      "realTimeFactor": 5.125979036816663,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.12639795243740082,
      "clippedFraction": 0,
      "rawPeak": 0.818787693977356,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 25090523136,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 309068.5188329996,
      "postInferenceRssBytes": 25045368832,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 1401.7276250001487,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 171733.389042,
      "durationMs": 171760.33966700017,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1746893843262072,
      "realTimeFactor": 5.7244463013999995,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 25016991744,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 173162.0672920003,
      "postInferenceRssBytes": 25017679872,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9979689717292786
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0012071100063621998
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003217258199583739
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000294676108751446
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00018586902297101915
        },
        {
          "label": "White noise and static.",
          "score": 0.00002160866279155016
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 594.1282079998018,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 171227.62541700012,
      "durationMs": 171254.311959,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1752053731221194,
      "realTimeFactor": 5.707587513900004,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11225259304046631,
      "clippedFraction": 0,
      "rawPeak": 0.9820544719696045,
      "rawClippedFraction": 0,
      "outputGain": 0.9979079857296672,
      "peakResidentMemoryBytes": 24999231488,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 171848.44016699982,
      "postInferenceRssBytes": 24999919616,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 64823.16708400049,
      "durationMs": 64860.111209000024,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4627975051130221,
      "realTimeFactor": 2.1607722361333495,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24804229120,
      "totalMs": 75192.82183399991,
      "postInferenceRssBytes": 1243381760,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 66233.46458299966,
      "durationMs": 66267.72012499988,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.4529432393258829,
      "realTimeFactor": 2.2077821527666552,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24804229120,
      "totalMs": 76600.43074999977,
      "postInferenceRssBytes": 1302937600,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9157536029815674
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.07782995700836182
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.005991353653371334
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00039727281546220183
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00002126855542883277
        },
        {
          "label": "White noise and static.",
          "score": 0.000006515838776977034
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 88510.34491700011,
      "durationMs": 88540.74395899988,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.33894343116764786,
      "realTimeFactor": 2.9503448305666704,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.12469441443681717,
      "clippedFraction": 0,
      "rawPeak": 1.1111327409744263,
      "rawClippedFraction": 0.000008680555555555556,
      "outputGain": 0.881982830548736,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24766480384,
      "totalMs": 98873.45458399977,
      "postInferenceRssBytes": 1309556736,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9974184036254883
        },
        {
          "label": "Solo piano music.",
          "score": 0.002452404238283634
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00007898049807408825
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00003342892887303606
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000009997223060054239
        },
        {
          "label": "White noise and static.",
          "score": 0.000006790255156374769
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 95964.20475000014,
      "durationMs": 95990.29420900025,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.31261656445915537,
      "realTimeFactor": 3.1988068250000046,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11531993746757507,
      "clippedFraction": 0,
      "rawPeak": 0.7632497549057007,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24785649664,
      "totalMs": 106323.00483400014,
      "postInferenceRssBytes": 1333444608,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9687485694885254
        },
        {
          "label": "Solo piano music.",
          "score": 0.02360290288925171
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.007393281441181898
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00019088336557615548
        },
        {
          "label": "White noise and static.",
          "score": 0.000036788449506275356
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000027526539270184003
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 90753.60912499946,
      "durationMs": 90781.60308299994,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.33056536582120394,
      "realTimeFactor": 3.025120304166649,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.15883134305477142,
      "clippedFraction": 0,
      "rawPeak": 1.1138813495635986,
      "rawClippedFraction": 0.0001420138888888889,
      "outputGain": 0.879806453698097,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24785649664,
      "totalMs": 101114.31370799983,
      "postInferenceRssBytes": 1340325888,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9986097812652588
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006553256534971297
        },
        {
          "label": "Solo piano music.",
          "score": 0.000592163298279047
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00008458585944026709
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00003298540468676947
        },
        {
          "label": "White noise and static.",
          "score": 0.000025127859771600924
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 88118.27795799945,
      "durationMs": 88148.26583299963,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.340451501041579,
      "realTimeFactor": 2.937275931933315,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.15203572809696198,
      "clippedFraction": 0,
      "rawPeak": 1.6546791791915894,
      "rawClippedFraction": 0.0004163194444444444,
      "outputGain": 0.5922598243357297,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24857788416,
      "totalMs": 98480.97645799952,
      "postInferenceRssBytes": 1367261184,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9988837838172913
        },
        {
          "label": "Solo piano music.",
          "score": 0.0007033791043795645
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00016773569223005325
        },
        {
          "label": "White noise and static.",
          "score": 0.00011710117541952059
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00010487096733413637
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000023134118237067014
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 93009.7889589997,
      "durationMs": 93040.19362500003,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3225466946626931,
      "realTimeFactor": 3.1003262986333238,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.10899408161640167,
      "clippedFraction": 0,
      "rawPeak": 2.2377727031707764,
      "rawClippedFraction": 0.007370138888888889,
      "outputGain": 0.437935451894379,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24857788416,
      "totalMs": 103372.90424999992,
      "postInferenceRssBytes": 1373700096,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85350.23958300008,
      "durationMs": 85374.74237500009,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3514928621943242,
      "realTimeFactor": 2.8450079861000024,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 95707.45299999998,
      "postInferenceRssBytes": 1380089856,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 84915.62637500011,
      "durationMs": 84945.90662500013,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3532918648861578,
      "realTimeFactor": 2.8305208791666705,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 95278.61725000002,
      "postInferenceRssBytes": 1386364928,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85314.36045800001,
      "durationMs": 85341.67066699956,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.35164068322083836,
      "realTimeFactor": 2.843812015266667,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24861982720,
      "totalMs": 95674.38129199945,
      "postInferenceRssBytes": 1396064256,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86636.2157499998,
      "durationMs": 86658.81570799957,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.34627551238582427,
      "realTimeFactor": 2.8878738583333265,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24899731456,
      "totalMs": 96991.52633299946,
      "postInferenceRssBytes": 1406500864,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85791.4590419996,
      "durationMs": 85819.03624999995,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.34968515904728187,
      "realTimeFactor": 2.8597153013999863,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24861982720,
      "totalMs": 96151.74687499984,
      "postInferenceRssBytes": 1417068544,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.8649240732192993
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.09604160487651825
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.03800676763057709
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0008581202710047364
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00008710977272130549
        },
        {
          "label": "White noise and static.",
          "score": 0.00008221629104809836
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 187036.10004199983,
      "durationMs": 187092.2972499993,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 0.32079368628049193,
      "realTimeFactor": 3.1172683340333305,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.12639792263507843,
      "clippedFraction": 0,
      "rawPeak": 0.8187971115112305,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 197425.0078749992,
      "postInferenceRssBytes": 1719844864,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 84634.99566599967,
      "durationMs": 84662.90987499997,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3544633016629535,
      "realTimeFactor": 2.821166522199989,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 94995.62049999986,
      "postInferenceRssBytes": 1733361664,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 84949.69074999972,
      "durationMs": 84979.58166700028,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3531501967239369,
      "realTimeFactor": 2.831656358333324,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 95312.29229200017,
      "postInferenceRssBytes": 1743945728,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86190.22120800037,
      "durationMs": 86216.88883300067,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3480673280510775,
      "realTimeFactor": 2.873007373600012,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 96549.59945800056,
      "postInferenceRssBytes": 1754480640,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86154.75591699942,
      "durationMs": 86183.80437499945,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.348210608697002,
      "realTimeFactor": 2.871825197233314,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 96516.51499999934,
      "postInferenceRssBytes": 1764966400,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85356.57287499998,
      "durationMs": 85382.10512499973,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3514667821063218,
      "realTimeFactor": 2.845219095833333,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 95714.81574999962,
      "postInferenceRssBytes": 1775517696,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85567.60850000047,
      "durationMs": 85590.88775000055,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3505999586280343,
      "realTimeFactor": 2.8522536166666823,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 95923.59837500044,
      "postInferenceRssBytes": 1786052608,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85310.25083400073,
      "durationMs": 85336.92550000068,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.35165762269735795,
      "realTimeFactor": 2.8436750278000242,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 95669.63612500057,
      "postInferenceRssBytes": 1796538368,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 85441.38733300043,
      "durationMs": 85471.15125000072,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3511178942246992,
      "realTimeFactor": 2.8480462444333474,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 95803.86187500061,
      "postInferenceRssBytes": 1807040512,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86179.32429100074,
      "durationMs": 86205.99437500004,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.34811133931265625,
      "realTimeFactor": 2.872644143033358,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25188941824,
      "totalMs": 96538.70499999993,
      "postInferenceRssBytes": 1817526272,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 10332.71062499989,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86269.25933299935,
      "durationMs": 86297.85962499955,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3477484359080909,
      "realTimeFactor": 2.8756419777666453,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4033691648,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 25226690560,
      "totalMs": 96630.57024999944,
      "postInferenceRssBytes": 1828012032,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 12756.374667000273,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 82585.16687500014,
      "durationMs": 82616.45062499974,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3632613595781385,
      "realTimeFactor": 2.752838895833338,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4028350464,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24804229120,
      "totalMs": 95372.82529200001,
      "postInferenceRssBytes": 1881194496,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9980049729347229
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0011915946379303932
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003125340736005455
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002891648036893457
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001804220664780587
        },
        {
          "label": "White noise and static.",
          "score": 0.000021232281142147258
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4812.832958999934,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 81995.30775000039,
      "durationMs": 82028.16475000055,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3658745948178951,
      "realTimeFactor": 2.733176925000013,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11225568503141403,
      "clippedFraction": 0,
      "rawPeak": 0.9819274544715881,
      "rawClippedFraction": 0,
      "outputGain": 0.9980370703937336,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24766480384,
      "totalMs": 86840.99770900048,
      "postInferenceRssBytes": 1906671616,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9976590871810913
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0016566356644034386
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00026408053236082196
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00025779171846807003
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0001524867839179933
        },
        {
          "label": "White noise and static.",
          "score": 0.000010005998774431646
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 560.779249999996,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 204638.87516700017,
      "durationMs": 204666.487167,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.1465997112010991,
      "realTimeFactor": 6.821295838900006,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.08269444108009338,
      "clippedFraction": 0,
      "rawPeak": 0.8592653274536133,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 3925.6579170000805,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 205227.266417,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 25419743232,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9663766026496887
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.02041863277554512
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.011952006258070469
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0010794500121846795
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00015341280959546566
        },
        {
          "label": "White noise and static.",
          "score": 0.00002000934364332352
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 560.779249999996,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 175144.51904099996,
      "durationMs": 175172.18941600004,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17128711857079143,
      "realTimeFactor": 5.838150634699999,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.11354653537273407,
      "clippedFraction": 0,
      "rawPeak": 0.9306381344795227,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4283.606250000048,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 175732.96866600003,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 25377685504,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9929277300834656
        },
        {
          "label": "Solo piano music.",
          "score": 0.0068979584611952305
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00011169482604600489
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00005566351683228277
        },
        {
          "label": "White noise and static.",
          "score": 0.000004221368726575747
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000002681727664821665
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 560.779249999996,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 175139.10099999976,
      "durationMs": 175165.42779099973,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17129241744823187,
      "realTimeFactor": 5.837970033333326,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06472428888082504,
      "clippedFraction": 0,
      "rawPeak": 0.6570315361022949,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4132.015375000265,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 175726.20704099972,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 25374932992,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9932330250740051
        },
        {
          "label": "Solo piano music.",
          "score": 0.0065468656830489635
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001817038719309494
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00002841012246790342
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000006967912213440286
        },
        {
          "label": "White noise and static.",
          "score": 0.000002978021029775846
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 560.779249999996,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 174928.73666700008,
      "durationMs": 174956.10479199968,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17149840884696355,
      "realTimeFactor": 5.830957888900002,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06733880937099457,
      "clippedFraction": 0,
      "rawPeak": 0.6881923079490662,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4178.240499999902,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 175516.88404199967,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 26171850752,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.6579445600509644
        },
        {
          "label": "Solo piano music.",
          "score": 0.31533846259117126
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.02300289086997509
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0031493499409407377
        },
        {
          "label": "White noise and static.",
          "score": 0.0004815920547116548
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00008316063031088561
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 560.779249999996,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 175071.481083,
      "durationMs": 175093.68845800008,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17135857773304175,
      "realTimeFactor": 5.8357160361,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.07768549770116806,
      "clippedFraction": 0,
      "rawPeak": 1.107345700263977,
      "rawClippedFraction": 0.0000017361111111111112,
      "outputGain": 0.884999146848523,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4051.58849999998,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 175654.46770800007,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 25374310400,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9069252610206604
        },
        {
          "label": "Solo piano music.",
          "score": 0.05621820688247681
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.03174125403165817
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0038592503406107426
        },
        {
          "label": "White noise and static.",
          "score": 0.0012047386262565851
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000051208786317147315
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 560.779249999996,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 174209.78454100032,
      "durationMs": 174235.08816599997,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.17220617130686763,
      "realTimeFactor": 5.806992818033344,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.0844251736998558,
      "clippedFraction": 0,
      "rawPeak": 1.0394583940505981,
      "rawClippedFraction": 3.4722222222222224e-7,
      "outputGain": 0.9427986782434854,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4014.7060829999646,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 174795.86741599996,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 25380618240,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.997639536857605
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0016766565386205912
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0002647706132847816
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0002573965466581285
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00015174136206042022
        },
        {
          "label": "White noise and static.",
          "score": 0.00000989995078271022
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4150.15512500031,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 83187.15808399975,
      "durationMs": 83218.6729999994,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3606325866993431,
      "realTimeFactor": 2.772905269466658,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.0826931893825531,
      "clippedFraction": 0,
      "rawPeak": 0.8592939376831055,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1735.0161249996745,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 28494610432,
      "totalMs": 87368.82812499971,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 2013495296,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.966738224029541
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.020119138062000275
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.011904766783118248
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0010667410679161549
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00015145672659855336
        },
        {
          "label": "White noise and static.",
          "score": 0.000019808454453595914
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4150.15512500031,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 84777.22387500035,
      "durationMs": 84801.3689999998,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3538686291996711,
      "realTimeFactor": 2.8259074625000116,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.11354649811983109,
      "clippedFraction": 0,
      "rawPeak": 0.9305417537689209,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1870.9534580002583,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31447400448,
      "totalMs": 88951.52412500011,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 2062450688,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9928956627845764
        },
        {
          "label": "Solo piano music.",
          "score": 0.0069305370561778545
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00011101931886514649
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0000558121464564465
        },
        {
          "label": "White noise and static.",
          "score": 0.000004253415227140067
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000027250453058513813
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4150.15512500031,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86222.25287500078,
      "durationMs": 86251.77370800066,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.34793802063478885,
      "realTimeFactor": 2.8740750958333594,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.0647270679473877,
      "clippedFraction": 0,
      "rawPeak": 0.6570295691490173,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1670.0928750005914,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31447400448,
      "totalMs": 90401.92883300097,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 2111733760,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9931772947311401
        },
        {
          "label": "Solo piano music.",
          "score": 0.006602135952562094
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0001824702339945361
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000028311000278335996
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000006879494321765378
        },
        {
          "label": "White noise and static.",
          "score": 0.0000029207333227532217
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4150.15512500031,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86104.81566700037,
      "durationMs": 86132.56016699961,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.3484125686537819,
      "realTimeFactor": 2.8701605222333457,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06734222173690796,
      "clippedFraction": 0,
      "rawPeak": 0.6881934404373169,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1831.1681670002145,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31409651712,
      "totalMs": 90282.71529199992,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 2160001024,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.6523499488830566
        },
        {
          "label": "Solo piano music.",
          "score": 0.3208957016468048
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.022907521575689316
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0032572841737419367
        },
        {
          "label": "White noise and static.",
          "score": 0.000506987445987761
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00008253745181718841
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4150.15512500031,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 87956.71779199984,
      "durationMs": 87984.22758400011,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.34107684726190035,
      "realTimeFactor": 2.931890593066661,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.07767856121063232,
      "clippedFraction": 0,
      "rawPeak": 1.1071666479110718,
      "rawClippedFraction": 0.0000017361111111111112,
      "outputGain": 0.8851422699997319,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1734.3287080002483,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31409651712,
      "totalMs": 92134.38270900042,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 2208841728,
      "fallback": false,
      "grade": "interactive"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9060094952583313
        },
        {
          "label": "Solo piano music.",
          "score": 0.05683610960841179
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.032001595944166183
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00387593568302691
        },
        {
          "label": "White noise and static.",
          "score": 0.0012239714851602912
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000052878312999382615
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4150.15512500031,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 86856.9166670004,
      "durationMs": 86883.49700000072,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 0.345395636308581,
      "realTimeFactor": 2.89523055556668,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08442602306604385,
      "clippedFraction": 0,
      "rawPeak": 1.0394480228424072,
      "rawClippedFraction": 3.4722222222222224e-7,
      "outputGain": 0.9428080851220972,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1721.638625000196,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 50,
      "sampling": {
        "shift": 1,
        "infer_steps": 50,
        "diffusion_guidance_scale": 7
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31447400448,
      "totalMs": 91033.65212500103,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d06de46b4622f781cf07f4a013a67d591ca52819",
      "postInferenceRssBytes": 2257174528,
      "fallback": false,
      "grade": "interactive"
    }
  }
]

ACE-Step 1.5 XL Turbo

52 configurations tested
Passed48
Failed4
TestedNov 7, 2026
Runtimeace-step-1.5 d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResultPlatformInference / loadTask metric
FP32mpsAcceleratedPassedFastmacos · aarch648976 ms / 13870 ms0.299 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648987 ms / 13870 ms0.300 RTF
FP32mpsAcceleratedPassedFastmacos · aarch648992 ms / 13870 ms0.300 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649024 ms / 13870 ms0.301 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649048 ms / 13870 ms0.302 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649057 ms / 13870 ms0.302 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649060 ms / 13870 ms0.302 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649102 ms / 13870 ms0.303 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649315 ms / 13870 ms0.311 RTF
FP32mpsAcceleratedPassedFastmacos · aarch649449 ms / 13870 ms0.315 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6410002 ms / 13870 ms0.333 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6420158 ms / 13870 ms0.336 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6410112 ms / 13870 ms0.337 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6410154 ms / 13870 ms0.338 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6410408 ms / 13870 ms0.347 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6410713 ms / 13870 ms0.357 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6410924 ms / 13870 ms0.364 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6411117 ms / 13870 ms0.371 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6411126 ms / 13870 ms0.371 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6411159 ms / 13870 ms0.372 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6411217 ms / 13870 ms0.374 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6411343 ms / 4294 ms0.378 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6411798 ms / 6934 ms0.393 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412018 ms / 4426 ms0.401 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412033 ms / 13870 ms0.401 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412086 ms / 4426 ms0.403 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412207 ms / 4426 ms0.407 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412388 ms / 4426 ms0.413 RTF
FP32mpsAcceleratedPassedFastmacos · aarch6412632 ms / 4426 ms0.421 RTF
FP32cpuPassedFastmacos · aarch6443302 ms / 587 ms0.722 RTF
FP32cpuPassedFastmacos · aarch6423521 ms / 587 ms0.784 RTF
FP32cpuPassedFastmacos · aarch6423575 ms / 587 ms0.786 RTF
FP32cpuPassedFastmacos · aarch6423581 ms / 587 ms0.786 RTF
FP32cpuPassedFastmacos · aarch6423693 ms / 587 ms0.790 RTF
FP32cpuPassedFastmacos · aarch6423693 ms / 587 ms0.790 RTF
FP32cpuPassedFastmacos · aarch6423740 ms / 587 ms0.791 RTF
FP32cpuPassedFastmacos · aarch6423761 ms / 587 ms0.792 RTF
FP32cpuPassedFastmacos · aarch6423946 ms / 587 ms0.798 RTF
FP32cpuPassedFastmacos · aarch6424250 ms / 587 ms0.808 RTF
FP32cpuPassedFastmacos · aarch6424392 ms / 587 ms0.813 RTF
FP32cpuPassedFastmacos · aarch6425595 ms / 576 ms0.853 RTF
FP32cpuPassedFastmacos · aarch6425631 ms / 581 ms0.854 RTF
FP32cpuPassedFastmacos · aarch6425759 ms / 587 ms0.859 RTF
FP32cpuPassedFastmacos · aarch6428692 ms / 551 ms0.956 RTF
FP32cpuPassedFastmacos · aarch6428711 ms / 551 ms0.957 RTF
FP32cpuPassedFastmacos · aarch6428854 ms / 551 ms0.962 RTF
FP32cpuPassedFastmacos · aarch6429398 ms / 551 ms0.980 RTF
FP32cpuPassedFastmacos · aarch6429825 ms / 551 ms0.994 RTF
FP32cpuFailed—macos · aarch6423583 ms / 587 ms0.786 RTF
FP32mpsFailed—macos · aarch648991 ms / 13870 ms0.300 RTF
FP32cpuFailed—macos · aarch6428799 ms / 551 ms0.960 RTF
FP32mpsFailed—macos · aarch6412233 ms / 4426 ms0.408 RTF
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-07T21:20:45.975901+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Torch2.10.0/Transformers4.57.6/Diffusers0.37.0/exact reviewed ACEd4a0b288b83ebb7e25a8c0b32c573c22e134e8ee", "interpreter": "App-owned CPython 3.11.18 private probe venv" }
Runtime
ace-step-1.5 d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee
Artifact revision
d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee
Fixture
{
  "file": "model-samples/desktop/ace-step-variants/xl-turbo/piano.txt",
  "size": 111,
  "sha1": "e774b12693cbe846aef04b5f4c19d4e548ff7268",
  "sha256": "1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea",
  "url": "https://assets.localai.tools/samples/v1/files/1c0a58ee3335db10628d7fd0f6aae93962029472c2a754c97c5b65d6ba271bea/piano.txt"
}
Rerun when
Checkpoint/runtime/adapter/config changes; Fixture or quality gate changes; Relevant device/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25759.108208003454,
      "durationMs": 25788.067708010203,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1646365921425372,
      "realTimeFactor": 0.8586369402667818,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25181143040,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 26375.02754101297,
      "postInferenceRssBytes": 25181863936,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23693.38987499941,
      "durationMs": 23720.16816699761,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.266175931695411,
      "realTimeFactor": 0.7897796624999803,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25263276032,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24307.128000000375,
      "postInferenceRssBytes": 25263292416,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9966849684715271
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0017772209830582142
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0009161821217276156
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0004710500652436167
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00013298798876348883
        },
        {
          "label": "White noise and static.",
          "score": 0.000017505170035292394
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23693.050834001042,
      "durationMs": 23715.666999996756,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.2661940503224718,
      "realTimeFactor": 0.789768361133368,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15707722306251526,
      "clippedFraction": 0,
      "rawPeak": 1.2089954614639282,
      "rawClippedFraction": 0.0014881944444444445,
      "outputGain": 0.8105903051226958,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24302.626832999522,
      "postInferenceRssBytes": 24897929216,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.961486279964447
        },
        {
          "label": "Solo piano music.",
          "score": 0.03798750415444374
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00040027592331171036
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0001108821015805006
        },
        {
          "label": "White noise and static.",
          "score": 0.000007833419658709317
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000007296719104488147
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23520.756124999025,
      "durationMs": 23544.89208299492,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.2754692000787557,
      "realTimeFactor": 0.7840252041666341,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15329115092754364,
      "clippedFraction": 0,
      "rawPeak": 0.99705570936203,
      "rawClippedFraction": 0,
      "outputGain": 0.9828939253826215,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24131.851915997686,
      "postInferenceRssBytes": 24809160704,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "fail",
    "quality": {
      "passed": false,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.502225935459137
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.2880725860595703
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.17820344865322113
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.030653638765215874
        },
        {
          "label": "White noise and static.",
          "score": 0.0006736432551406324
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0001708062191028148
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23583.196874998976,
      "durationMs": 23610.517166991485,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.2720921662577311,
      "realTimeFactor": 0.7861065624999658,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.245652437210083,
      "clippedFraction": 0,
      "rawPeak": 1.1572144031524658,
      "rawClippedFraction": 0.0007982638888888889,
      "outputGain": 0.8468612189152667,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24197.47699999425,
      "postInferenceRssBytes": 24891293696,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9952228665351868
        },
        {
          "label": "Solo piano music.",
          "score": 0.004033244214951992
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0003917203866876662
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00023365409288089722
        },
        {
          "label": "White noise and static.",
          "score": 0.00008469721797155216
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00003394286977709271
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23739.626416994724,
      "durationMs": 23765.835458005313,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.26370986101633,
      "realTimeFactor": 0.7913208805664909,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15500876307487488,
      "clippedFraction": 0,
      "rawPeak": 1.9644818305969238,
      "rawClippedFraction": 0.0035256944444444446,
      "outputGain": 0.49885928428374365,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24352.79529100808,
      "postInferenceRssBytes": 24961826816,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9992250204086304
        },
        {
          "label": "Solo piano music.",
          "score": 0.0003296384238637984
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00023286622308660299
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0001303144235862419
        },
        {
          "label": "White noise and static.",
          "score": 0.00006825952732469887
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000013900800695409998
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23760.768291002023,
      "durationMs": 23786.883875000058,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.2625854363202857,
      "realTimeFactor": 0.7920256097000674,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.20921966433525085,
      "clippedFraction": 0,
      "rawPeak": 1.677140474319458,
      "rawClippedFraction": 0.003695138888888889,
      "outputGain": 0.5843279170742449,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24373.843708002823,
      "postInferenceRssBytes": 24830214144,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 24250.332834009896,
      "durationMs": 24278.136624998297,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.237096422772659,
      "realTimeFactor": 0.8083444278003299,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24865.096458001062,
      "postInferenceRssBytes": 24954159104,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 24392.22762499412,
      "durationMs": 24417.487041995628,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.2298999690073298,
      "realTimeFactor": 0.8130742541664707,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 25004.446874998393,
      "postInferenceRssBytes": 24844730368,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23945.723208002164,
      "durationMs": 23972.540000002482,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.2528333239053988,
      "realTimeFactor": 0.7981907736000722,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24559.499833005248,
      "postInferenceRssBytes": 24915197952,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23581.444916009787,
      "durationMs": 23607.32129099779,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.272186675025692,
      "realTimeFactor": 0.7860481638669928,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24194.281124000554,
      "postInferenceRssBytes": 24920965120,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 23575.030292005977,
      "durationMs": 23601.660084008472,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.27253282937128,
      "realTimeFactor": 0.7858343430668658,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 24188.619917011238,
      "postInferenceRssBytes": 24926732288,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9939947724342346
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0033683779183775187
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0018311351304873824
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.000742014788556844
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000049537495215190575
        },
        {
          "label": "White noise and static.",
          "score": 0.000014183339771989267
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 586.9598330027657,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 43301.87883399776,
      "durationMs": 43357.04583399638,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 1.3856211696960359,
      "realTimeFactor": 0.7216979805666294,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.14953109622001648,
      "clippedFraction": 0,
      "rawPeak": 1.0765608549118042,
      "rawClippedFraction": 0.000030555555555555554,
      "outputGain": 0.9103061805830616,
      "peakResidentMemoryBytes": 25265389568,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 43944.005666999146,
      "postInferenceRssBytes": 25065504768,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 580.6942079943838,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25631.014208003762,
      "durationMs": 25656.774957999005,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.170457000122607,
      "realTimeFactor": 0.8543671402667921,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25179144192,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 26237.46916599339,
      "postInferenceRssBytes": 25179832320,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9981974959373474
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006977921002544463
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005314565496519208
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00039151660166680813
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016436849546153098
        },
        {
          "label": "White noise and static.",
          "score": 0.000017431157175451517
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 575.8900829969207,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 25595.10112499993,
      "durationMs": 25620.93966700195,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.1720992956225362,
      "realTimeFactor": 0.8531700374999976,
      "activationPlacement": {
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15474623441696167,
      "clippedFraction": 0,
      "rawPeak": 1.0675406455993652,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179978336560516,
      "peakResidentMemoryBytes": 25180602368,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 26196.82974999887,
      "postInferenceRssBytes": 25184141312,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 11159.442208998371,
      "durationMs": 11192.551583997556,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.6883064079860213,
      "realTimeFactor": 0.37198140696661236,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24725733376,
      "totalMs": 25062.881876001484,
      "postInferenceRssBytes": 1212284928,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9048.193666996667,
      "durationMs": 9079.088500002399,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.3155788994023365,
      "realTimeFactor": 0.3016064555665556,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24725733376,
      "totalMs": 22949.418792006327,
      "postInferenceRssBytes": 1278410752,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9966495633125305
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0018017884576693177
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.000917074445169419
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00048002187395468354
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00013400401803664863
        },
        {
          "label": "White noise and static.",
          "score": 0.00001766086097632069
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8976.027958997292,
      "durationMs": 9003.831209003692,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.3422355787037104,
      "realTimeFactor": 0.29920093196657643,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.15707668662071228,
      "clippedFraction": 0,
      "rawPeak": 1.2089948654174805,
      "rawClippedFraction": 0.0014875,
      "outputGain": 0.8105907047517478,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24725733376,
      "totalMs": 22874.16150100762,
      "postInferenceRssBytes": 1284112384,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9612339735031128
        },
        {
          "label": "Solo piano music.",
          "score": 0.038231126964092255
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0004069726273883134
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00011266893852734938
        },
        {
          "label": "White noise and static.",
          "score": 0.000007858477147237863
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000073069686550297774
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9059.606834009173,
      "durationMs": 9083.559625010821,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.311401979099353,
      "realTimeFactor": 0.30198689446697247,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1532931625843048,
      "clippedFraction": 0,
      "rawPeak": 0.9970396757125854,
      "rawClippedFraction": 0,
      "outputGain": 0.9829097315506454,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24736120832,
      "totalMs": 22953.88991701475,
      "postInferenceRssBytes": 1304674304,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "fail",
    "quality": {
      "passed": false,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.501242458820343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.2888043224811554
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.17834074795246124
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.030764682218432426
        },
        {
          "label": "White noise and static.",
          "score": 0.0006761247641406953
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00017175216635223478
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8991.367709008045,
      "durationMs": 9017.851584008895,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.3365335476097093,
      "realTimeFactor": 0.2997122569669349,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.2456507384777069,
      "clippedFraction": 0,
      "rawPeak": 1.1572173833847046,
      "rawClippedFraction": 0.0007982638888888889,
      "outputGain": 0.8468590379567513,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24736120832,
      "totalMs": 22888.181876012823,
      "postInferenceRssBytes": 1307705344,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9952400922775269
        },
        {
          "label": "Solo piano music.",
          "score": 0.004015402402728796
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00039126735646277666
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.00023431843146681786
        },
        {
          "label": "White noise and static.",
          "score": 0.00008472202898701653
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00003416507752262987
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9102.477667009225,
      "durationMs": 9128.205583998351,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.2958059439938197,
      "realTimeFactor": 0.30341592223364083,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.15501008927822113,
      "clippedFraction": 0,
      "rawPeak": 1.9644733667373657,
      "rawClippedFraction": 0.0035256944444444446,
      "outputGain": 0.498861433600193,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24764776448,
      "totalMs": 22998.53587600228,
      "postInferenceRssBytes": 1330642944,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.9992259740829468
        },
        {
          "label": "Solo piano music.",
          "score": 0.00032886440749280155
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00023269660596270114
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0001304048055317253
        },
        {
          "label": "White noise and static.",
          "score": 0.00006817267421865836
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000013909207154938485
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9023.674374999246,
      "durationMs": 9048.650041993824,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.324588050641234,
      "realTimeFactor": 0.3007891458333082,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.20921674370765686,
      "clippedFraction": 0,
      "rawPeak": 1.6771618127822876,
      "rawClippedFraction": 0.0036947916666666665,
      "outputGain": 0.5843204826934691,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24764776448,
      "totalMs": 22918.98033399775,
      "postInferenceRssBytes": 1333346304,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8992.145709009492,
      "durationMs": 9018.442959000822,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.3362448708923975,
      "realTimeFactor": 0.2997381903003164,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24768970752,
      "totalMs": 22888.77325100475,
      "postInferenceRssBytes": 1335836672,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 8986.593292007456,
      "durationMs": 9010.807166996528,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.3383061884731737,
      "realTimeFactor": 0.29955310973358185,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24768970752,
      "totalMs": 22881.137459000456,
      "postInferenceRssBytes": 1338343424,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9057.497250003507,
      "durationMs": 9083.48775000195,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.3121732385829192,
      "realTimeFactor": 0.3019165750001169,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24768970752,
      "totalMs": 22953.818042005878,
      "postInferenceRssBytes": 1343750144,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9315.240959011135,
      "durationMs": 9347.881292007514,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.2205286081171507,
      "realTimeFactor": 0.3105080319670378,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24768970752,
      "totalMs": 23218.211584011442,
      "postInferenceRssBytes": 1350533120,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 9449.006082999404,
      "durationMs": 9476.963542008889,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 3.174937103064821,
      "realTimeFactor": 0.3149668694333135,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24768970752,
      "totalMs": 23347.293834012817,
      "postInferenceRssBytes": 1357201408,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.99396812915802
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.003391522914171219
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.001825304701924324
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0007509997813031077
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000049710604798747227
        },
        {
          "label": "White noise and static.",
          "score": 0.000014360151908476837
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 20158.026707998943,
      "durationMs": 20218.85604200361,
      "audioSeconds": 60,
      "audioSecondsPerSecond": 2.97648181883752,
      "realTimeFactor": 0.3359671117999824,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1495276391506195,
      "clippedFraction": 0,
      "rawPeak": 1.0765800476074219,
      "rawClippedFraction": 0.000030555555555555554,
      "outputGain": 0.9102899521293747,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24896323584,
      "totalMs": 34089.186334007536,
      "postInferenceRssBytes": 1642790912,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 10002.133083995432,
      "durationMs": 10028.694874999928,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.9993602112736797,
      "realTimeFactor": 0.33340443613318105,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 23899.025167003856,
      "postInferenceRssBytes": 1652523008,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 10112.276500003645,
      "durationMs": 10137.284875003388,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.966691031439774,
      "realTimeFactor": 0.33707588333345484,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 24007.615167007316,
      "postInferenceRssBytes": 1659256832,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 10154.004625001107,
      "durationMs": 10180.684833001578,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.9544993436514932,
      "realTimeFactor": 0.33846682083337026,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 24051.015125005506,
      "postInferenceRssBytes": 1665941504,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 10408.264083001995,
      "durationMs": 10435.624332996667,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.882325021805872,
      "realTimeFactor": 0.3469421361000665,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 24305.954625000595,
      "postInferenceRssBytes": 1672544256,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 10712.652332993457,
      "durationMs": 10736.154707992682,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.800426922061517,
      "realTimeFactor": 0.3570884110997819,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 24606.48499999661,
      "postInferenceRssBytes": 1679261696,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 11216.916124991258,
      "durationMs": 11246.901374994195,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.674531900364315,
      "realTimeFactor": 0.37389720416637523,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24896323584,
      "totalMs": 25117.231666998123,
      "postInferenceRssBytes": 1685962752,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 11117.421792005189,
      "durationMs": 11145.318417009548,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.6984673750143884,
      "realTimeFactor": 0.370580726400173,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 25015.648709013476,
      "postInferenceRssBytes": 1692237824,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 10924.096250004368,
      "durationMs": 10947.497542001656,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.746222599419884,
      "realTimeFactor": 0.36413654166681225,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 24817.827834005584,
      "postInferenceRssBytes": 1698938880,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 11126.23516698659,
      "durationMs": 11152.003666997189,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.6963298501019506,
      "realTimeFactor": 0.37087450556621965,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 25022.333959001116,
      "postInferenceRssBytes": 1705590784,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 13870.330292003928,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12032.58154199284,
      "durationMs": 12059.821292001288,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.4932305586546133,
      "realTimeFactor": 0.40108605139976133,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4030873600,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24934072320,
      "totalMs": 25930.151584005216,
      "postInferenceRssBytes": 1712226304,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 6934.167500003241,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 11798.143333988264,
      "durationMs": 11831.923166988418,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.542772972894435,
      "realTimeFactor": 0.3932714444662755,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4042588160,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24725733376,
      "totalMs": 18766.09066699166,
      "postInferenceRssBytes": 1873608704,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9982074499130249
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0007111236336641014
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0005166306509636343
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00038522566319443285
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00016278441762551665
        },
        {
          "label": "White noise and static.",
          "score": 0.00001683093978499528
        }
      ],
      "metric": "Frozen CLAP top1 among six labels plus delivered signal gates"
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4293.958583002677,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 11342.769916998805,
      "durationMs": 11373.793207996641,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.6448566108213654,
      "realTimeFactor": 0.37809233056662683,
      "activationPlacement": {
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1547452211380005,
      "clippedFraction": 0,
      "rawPeak": 1.0675445795059204,
      "rawClippedFraction": 0.000024652777777777778,
      "outputGain": 0.9179944508299244,
      "peakResidentMemoryBytes": 4043374592,
      "memoryMeasurementScope": "Process lifetime peak RSS through complete WAV construction, not a validated minimum or system-wide memory measurement.",
      "referenceInput": null,
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 24725733376,
      "totalMs": 15667.751790999318,
      "postInferenceRssBytes": 1872478208,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.998995840549469
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006129964021965861
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002626303758006543
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00008229568629758433
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000043573272705543786
        },
        {
          "label": "White noise and static.",
          "score": 0.0000027568087261897745
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 550.6867919903016,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 29824.996791998274,
      "durationMs": 29852.088416999322,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0058676689631254,
      "realTimeFactor": 0.9941665597332758,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.15972521901130676,
      "clippedFraction": 0,
      "rawPeak": 1.0238733291625977,
      "rawClippedFraction": 0.000005208333333333333,
      "outputGain": 0.9571496513162613,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4138.353874994209,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 30402.775208989624,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 35466657792,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9887890219688416
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.0060493857599794865
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002781439805403352
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0020346653182059526
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0003282790712546557
        },
        {
          "label": "White noise and static.",
          "score": 0.000017264817870454863
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 550.6867919903016,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 29397.763165994547,
      "durationMs": 29422.400082999957,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0204858046717677,
      "realTimeFactor": 0.9799254388664849,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.17141057550907135,
      "clippedFraction": 0,
      "rawPeak": 1.1202718019485474,
      "rawClippedFraction": 0.0002989583333333333,
      "outputGain": 0.8747877062472114,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "817e0639794f2e7337f569f5470117d0a127829d534e64f24f6078843b02a222",
        "encodeMs": 4881.686167005682,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 29973.08687499026,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 29413408768,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9367057085037231
        },
        {
          "label": "Solo piano music.",
          "score": 0.06302202492952347
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.00013632443733513355
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00012991308176424354
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.0000041452899495197926
        },
        {
          "label": "White noise and static.",
          "score": 0.0000018536833295001998
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 550.6867919903016,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 28853.88445800345,
      "durationMs": 28878.714042002684,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0397213603480222,
      "realTimeFactor": 0.961796148600115,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.05833323672413826,
      "clippedFraction": 0,
      "rawPeak": 0.642917811870575,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4349.185958999442,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 29429.400833992986,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 30605361152,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.6182007193565369
        },
        {
          "label": "Solo piano music.",
          "score": 0.37607431411743164
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004547132644802332
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0011559497797861695
        },
        {
          "label": "White noise and static.",
          "score": 0.000012906315532745793
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000009007019798445981
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 550.6867919903016,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 28691.643333004322,
      "durationMs": 28718.791582999984,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0456006179851909,
      "realTimeFactor": 0.9563881111001441,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.08131368458271027,
      "clippedFraction": 0,
      "rawPeak": 0.7587364315986633,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "11e1aa82c221570c64c06193423f8e24ac6082f8632ca71ba4a526a4b94c6e69",
        "encodeMs": 4285.686375005753,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 29269.478374990285,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 29081518080,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "fail",
    "quality": {
      "passed": false,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.6265242099761963
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.3115238845348358
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.056648049503564835
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.004977040458470583
        },
        {
          "label": "White noise and static.",
          "score": 0.00021440288401208818
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011240153980907053
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 550.6867919903016,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 28799.334249997628,
      "durationMs": 28826.86499999545,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.041690746722816,
      "realTimeFactor": 0.9599778083332543,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.06652025878429413,
      "clippedFraction": 0,
      "rawPeak": 1.5594267845153809,
      "rawClippedFraction": 0.00013298611111111112,
      "outputGain": 0.6284360444049651,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4368.161833990598,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 29377.551791985752,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 28879372288,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.5216854810714722
        },
        {
          "label": "Solo piano music.",
          "score": 0.4321214258670807
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0415525883436203
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.004381605423986912
        },
        {
          "label": "White noise and static.",
          "score": 0.00013102118100505322
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012780081306118518
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 550.6867919903016,
      "placement": {
        "diffusion": {
          "cpu/torch.float32": 829
        },
        "textEncoder": {
          "cpu/torch.float32": 310
        },
        "audioDecoder": {
          "cpu/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 28710.97537499736,
      "durationMs": 28736.559166994994,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 1.0448965807732598,
      "realTimeFactor": 0.9570325124999121,
      "activationPlacement": {
        "referenceEncoder": "cpu",
        "text": "cpu",
        "diffusion": "cpu",
        "vae": "cpu"
      },
      "waveformDevice": "cpu",
      "rms": 0.08430834114551544,
      "clippedFraction": 0,
      "rawPeak": 1.227309226989746,
      "rawClippedFraction": 0.0000010416666666666667,
      "outputGain": 0.7984947708766698,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "57a8ead2921a6dbc6286fedbeb2219b9238bc705590ef383927edeb232ced363",
        "encodeMs": 4219.204499997431,
        "posteriorSeed": 0,
        "encoderDevice": "cpu",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": null,
      "totalMs": 29287.245958985295,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 28055175168,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.998984158039093
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0006249137222766876
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0002619844162836671
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00008265945507446304
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000043452586396597326
        },
        {
          "label": "White noise and static.",
          "score": 0.000002765199496934656
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4425.977209000848,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12632.001291989582,
      "durationMs": 12664.20516699145,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.374920593067395,
      "realTimeFactor": 0.42106670973298604,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.1597135215997696,
      "clippedFraction": 0,
      "rawPeak": 1.0239264965057373,
      "rawClippedFraction": 0.000005208333333333333,
      "outputGain": 0.9570999513581869,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1608.4085840120679,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 28412493824,
      "totalMs": 17090.1823759923,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 1993031680,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo piano music.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.9888652563095093
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.006007404997944832
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.002752046100795269
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0020307879894971848
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00032743546762503684
        },
        {
          "label": "White noise and static.",
          "score": 0.0000171306892298162
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4425.977209000848,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12388.06866700179,
      "durationMs": 12415.09741700429,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.421684994361653,
      "realTimeFactor": 0.412935622233393,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.17140214145183563,
      "clippedFraction": 0,
      "rawPeak": 1.120309829711914,
      "rawClippedFraction": 0.0002986111111111111,
      "outputGain": 0.8747580124794634,
      "referenceInput": {
        "inputSha256": "7e436e48a54de63ff8c32be9d4538afbfdd2954d3a48151c7244bf639171d847",
        "latentSha256": "70a3a09c0ebe66416b75d66f6d68f643ecf1e6789c32f5ba39bcecf36585b4ea",
        "encodeMs": 1673.9116250100778,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31365283840,
      "totalMs": 16841.07462600514,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 2039316480,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.9357886910438538
        },
        {
          "label": "Solo piano music.",
          "score": 0.06394156068563461
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0001358777954010293
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.00012801391130778939
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000004127711235923925
        },
        {
          "label": "White noise and static.",
          "score": 0.0000018010600797424559
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4425.977209000848,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12206.847791996552,
      "durationMs": 12232.190708004055,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.457636935529709,
      "realTimeFactor": 0.40689492639988506,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.05833512917160988,
      "clippedFraction": 0,
      "rawPeak": 0.6429407596588135,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1548.7065420020372,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31365283840,
      "totalMs": 16658.167917004903,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 2084012032,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Solo acoustic guitar music.",
      "predictions": [
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.6179400682449341
        },
        {
          "label": "Solo piano music.",
          "score": 0.3764384388923645
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.004466279409825802
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.0011333947768434882
        },
        {
          "label": "White noise and static.",
          "score": 0.000012673061064560898
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.000009143894203589298
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4425.977209000848,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12086.307999998098,
      "durationMs": 12114.543709001737,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.4821475673137505,
      "realTimeFactor": 0.40287693333326996,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08131671696901321,
      "clippedFraction": 0,
      "rawPeak": 0.7589317560195923,
      "rawClippedFraction": 0,
      "outputGain": 1,
      "referenceInput": {
        "inputSha256": "e7f487391b34fc995bcca221eaceb2f46b046fce20d995fe65af849bceb21421",
        "latentSha256": "239350ae01ffb1b87198e8699f592a7519db0b4c4e464335fa53d9f7ffe7797a",
        "encodeMs": 1500.718292008969,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31365283840,
      "totalMs": 16540.520918002585,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 2128625664,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "fail",
    "quality": {
      "passed": false,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Solo piano music.",
          "score": 0.6243315935134888
        },
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.3136337697505951
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.0566893108189106
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.005016625393182039
        },
        {
          "label": "White noise and static.",
          "score": 0.00021638614998664707
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00011229686060687527
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4425.977209000848,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12232.826416002354,
      "durationMs": 12259.184875001665,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.4524176980681704,
      "realTimeFactor": 0.4077608805334118,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.06652122735977173,
      "clippedFraction": 0,
      "rawPeak": 1.559402585029602,
      "rawClippedFraction": 0.00013298611111111112,
      "outputGain": 0.628445796748116,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1532.1964589966228,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 42,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31365283840,
      "totalMs": 16685.162084002513,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 2173648896,
      "fallback": false,
      "grade": "fast"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "expected": "Electronic dance music with drums and synthesizers.",
      "predictions": [
        {
          "label": "Electronic dance music with drums and synthesizers.",
          "score": 0.5162053108215332
        },
        {
          "label": "Solo piano music.",
          "score": 0.437502384185791
        },
        {
          "label": "Solo acoustic guitar music.",
          "score": 0.04167938232421875
        },
        {
          "label": "Rain falling and water splashing.",
          "score": 0.004355120938271284
        },
        {
          "label": "White noise and static.",
          "score": 0.0001318583672400564
        },
        {
          "label": "A person giving a spoken speech.",
          "score": 0.00012594136933330446
        }
      ],
      "metric": "Frozen reference-instrument/style proxy; no broad human timbre fidelity claim."
    },
    "runtime": {
      "engine": "ace-step-1.5",
      "version": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "torchVersion": "2.10.0",
      "transformersVersion": "4.57.6",
      "diffusersVersion": "0.37.0",
      "loadMs": 4425.977209000848,
      "placement": {
        "diffusion": {
          "mps:0/torch.float32": 829
        },
        "textEncoder": {
          "mps:0/torch.float32": 310
        },
        "audioDecoder": {
          "mps:0/torch.float32": 365
        }
      },
      "parameterDtype": "float32",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "accelerationDetails": "All neural components execute on the selected device; tokenization, chunk assembly and WAV encoding use CPU.",
      "inferenceMs": 12018.39591600583,
      "durationMs": 12044.336165999994,
      "audioSeconds": 30,
      "audioSecondsPerSecond": 2.4961733836748277,
      "realTimeFactor": 0.40061319720019434,
      "activationPlacement": {
        "referenceEncoder": "mps:0",
        "text": "mps:0",
        "diffusion": "mps:0",
        "vae": "mps:0"
      },
      "waveformDevice": "mps:0",
      "rms": 0.08432228863239288,
      "clippedFraction": 0,
      "rawPeak": 1.2271027565002441,
      "rawClippedFraction": 0.0000010416666666666667,
      "outputGain": 0.7986291244223156,
      "referenceInput": {
        "inputSha256": "21ded90a569b38f36dd233ba273d0d43149b8e5fc63240d5570a020249a81c49",
        "latentSha256": "71aec06f2915030185e33dbefc5bd9171c2e88f8381dd9f1f53fcfed1ec8e2c8",
        "encodeMs": 1600.5853339884197,
        "posteriorSeed": 0,
        "encoderDevice": "mps:0",
        "originalShape": [
          1,
          2,
          1440000
        ]
      },
      "seed": 123,
      "steps": 8,
      "sampling": {
        "fix_nfe": 8,
        "shift": 3
      },
      "planningLm": false,
      "noiseGenerator": "cpu",
      "driverAllocatedMemoryBytes": 31365283840,
      "totalMs": 16470.313375000842,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 8,
      "artifactRevision": "d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee",
      "postInferenceRssBytes": 2219311104,
      "fallback": false,
      "grade": "fast"
    }
  }
]

Something not quite right?

Tell us what happened with ACE-Step 1.5. Loading, results, or the page itself—we’re here for all of it.

Attachments Optional

Up to 3 files · 50 MB each

Browser, OS, CPU/GPU, memory, screen, storage, and network details are included automatically. Model inputs, results, and browsing history are not collected.