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.

Grounding DINO Desktop

Find objects in a still image using English descriptions, returning confidence scores and pixel-coordinate boxes through native Transformers.

Desktop inferencePrivate · Runs on your computer
About this model

Find objects in a still image using English descriptions, returning confidence scores and pixel-coordinate boxes through native Transformers.

Created by
IDEA Research
License
Apache-2.0
Model released
Unknown
Source repository created
Sep 25, 2023
Model weights · FP32
933.4 MB

Best for

  • Local image object localization
  • Interactive visual exploration with human review

Keep in mind

  • Can confidently detect absent objects, miss real objects and merge/truncate phrases; confidence is not a calibrated probability of correctness.
  • Unnormalized article-heavy prompts failed the absent-airplane quality gate on the dog image across CPU/GPU and two Transformers versions. Default leading-article removal and0.4 threshold pass this small suite; these changes do not guarantee absent-object rejection generally. Original failures are retained.
  • Default label normalization lowercases, removes a leading English a/an/the and appends periods. Disable normalizeLabels to use literal labels; that path is not quality certified. Prefer short English noun descriptions.
  • Three prompt/image cases on two photos are compatibility checks, not COCO/OpenImages accuracy or general open-vocabulary validation.
  • Accepts still JPEG/PNG, applies EXIF orientation and composites transparency over white. Official processor resizes internally; original pixel-coordinate output is retained.
  • No video tracking, segmentation, streaming or safety-critical detection support.
  • The app downloads its own pinned CPython and isolated packages on request; installed Python, user packages and Python/pip configuration are not used. Automatic model-package provisioning is verified on Apple Silicon Mac; other platform stacks remain 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/IDEA-Research/grounding-dino-base
Release verification
Unknown
Release evidence
Exact original checkpoint public release date has not been verified. Repository creation, training, paper, artifact conversion and update dates are not release evidence.
Source updated
May 12, 2024
Native model files
https://huggingface.co/IDEA-Research/grounding-dino-base
Files updated
May 12, 2024
Artifact revision
12bdfa3120f3e7ec7b434d90674b3396eccf88eb

Desktop app

Sample result

Grounding DINO Base — Object Detection Desktop

Run Grounding DINO Desktop 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 / FP32Tested16.0 GiBRAM estimate
macOS Apple SiliconMPS / BF16Tested16.0 GiBRAM estimate
macOS Apple SiliconMPS / FP16Tested16.0 GiBRAM estimate
macOS Apple SiliconMPS / FP32Tested16.0 GiBRAM estimate
Windows Intel / AMDCPU / FP32Unverified16.0 GiBRAM estimate
macOS Intel / AMDCPU / FP32Unverified16.0 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.

Grounding DINO Base — Object Detection Desktop

64 configurations tested
Passed64
Failed0
TestedNov 1, 2026
Runtimetransformers 5.22.1

Quality criteria

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

PrecisionDeviceResultPlatformInference / load
BF16mpsAcceleratedPassedNot measuredmacos · aarch64406 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64635 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch641893 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64409 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64401 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64405 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64424 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64406 ms / 2269 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64412 ms / 2412 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64740 ms / 2412 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64536 ms / 2412 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64414 ms / 2319 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64668 ms / 2319 ms
BF16mpsAcceleratedPassedNot measuredmacos · aarch64489 ms / 2319 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64419 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64640 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch641972 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64405 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64423 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64409 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64412 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64400 ms / 2527 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64392 ms / 2278 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64642 ms / 2278 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64477 ms / 2278 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64391 ms / 2338 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64637 ms / 2338 ms
FP16mpsAcceleratedPassedNot measuredmacos · aarch64466 ms / 2338 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64480 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64741 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64512 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64474 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64465 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64458 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64462 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64458 ms / 2538 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64447 ms / 2390 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64687 ms / 2390 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64515 ms / 2390 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64457 ms / 2372 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64689 ms / 2372 ms
FP32mpsAcceleratedPassedNot measuredmacos · aarch64505 ms / 2372 ms
FP32cpuPassedNot measuredmacos · aarch641919 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch642412 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch641955 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch641896 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch641879 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch642055 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch642227 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch641996 ms / 1992 ms
FP32cpuPassedNot measuredmacos · aarch641420 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641684 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641406 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641363 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641342 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641336 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641431 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641489 ms / 1932 ms
FP32cpuPassedNot measuredmacos · aarch641373 ms / 1949 ms
FP32cpuPassedNot measuredmacos · aarch641682 ms / 1949 ms
FP32cpuPassedNot measuredmacos · aarch641373 ms / 1949 ms
FP32cpuPassedNot measuredmacos · aarch641385 ms / 2012 ms
FP32cpuPassedNot measuredmacos · aarch641803 ms / 2012 ms
FP32cpuPassedNot measuredmacos · aarch641378 ms / 2012 ms
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-01T14:59:01.886969+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2" }
Runtime
transformers 5.22.1
Artifact revision
12bdfa3120f3e7ec7b434d90674b3396eccf88eb
Fixture
{
  "file": "model-samples/desktop/mobilenet-dog.jpg",
  "size": 83281,
  "sha1": "f1d042827c2f2154a6d8043b8e8f07c1be612941",
  "sha256": "d4cb12ea8a768582562e1f1c0c930ea6fc4757f5b66551b986b41498efd75ca1",
  "url": "https://assets.localai.tools/samples/v1/files/d4cb12ea8a768582562e1f1c0c930ea6fc4757f5b66551b986b41498efd75ca1/mobilenet-dog.jpg"
}
Rerun when
artifact/runtime/processor/backend/driver; fixture checksum/reference box/prompt; thresholds and label normalization; quality gates
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1923.5736669998005,
      "inferenceMs": 1919.4210000005114,
      "imagesPerSecond": 0.5209904445141184,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3925.5994579998514,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.741491350405051,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 2417.721125000753,
      "inferenceMs": 2412.2006249999686,
      "imagesPerSecond": 0.41455921602707196,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 4420.432791999701,
      "grade": "C"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1959.5026669994695,
      "inferenceMs": 1955.2334589989186,
      "imagesPerSecond": 0.5114478761589938,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3961.5360830011923,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1899.7544580015528,
      "inferenceMs": 1895.6222090000665,
      "imagesPerSecond": 0.5275312745610299,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3901.4297499998065,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1882.726707999609,
      "inferenceMs": 1878.5421659995336,
      "imagesPerSecond": 0.5323276837216591,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3884.6378329999425,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 2058.693041999504,
      "inferenceMs": 2054.598917000476,
      "imagesPerSecond": 0.48671299869071644,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 4060.3701239997463,
      "grade": "C"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 2232.4927499994374,
      "inferenceMs": 2227.3354580011073,
      "imagesPerSecond": 0.44896694676491916,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 4235.914041000797,
      "grade": "C"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1992.311707999761,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 2000.0104579994513,
      "inferenceMs": 1995.7772919988201,
      "imagesPerSecond": 0.5010579106241235,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 4001.8591249991005,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1423.3994589994836,
      "inferenceMs": 1420.3199999992648,
      "imagesPerSecond": 0.7040666891971652,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3364.815457998702,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.741491350405051,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1687.2268749993964,
      "inferenceMs": 1683.6594169999444,
      "imagesPerSecond": 0.5939443511573534,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3629.135123997912,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1408.569749999515,
      "inferenceMs": 1405.5330409992166,
      "imagesPerSecond": 0.7114738471669677,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3350.2153739991627,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1366.3829999986774,
      "inferenceMs": 1363.3223749984609,
      "imagesPerSecond": 0.7335022283347539,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3307.572583000365,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1345.3709170007642,
      "inferenceMs": 1342.2932920002495,
      "imagesPerSecond": 0.7449936656614194,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3286.564415999237,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1339.4559580010537,
      "inferenceMs": 1336.3623330005794,
      "imagesPerSecond": 0.7483000495491865,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3280.7101249982225,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1433.9341670001886,
      "inferenceMs": 1430.8130420013185,
      "imagesPerSecond": 0.6989033302361235,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3375.4719579992525,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1931.872207998822,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1492.528291999406,
      "inferenceMs": 1489.4747500002268,
      "imagesPerSecond": 0.6713776114699814,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3433.8297079975746,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1949.2624159993284,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1375.9487079987593,
      "inferenceMs": 1372.945584000263,
      "imagesPerSecond": 0.7283609865194836,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3334.6955409997463,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.741491350405051,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1949.2624159993284,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1686.1683749993972,
      "inferenceMs": 1682.4417500010895,
      "imagesPerSecond": 0.5943742183046471,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3645.682082998974,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 1949.2624159993284,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1376.4461669998127,
      "inferenceMs": 1373.488583998551,
      "imagesPerSecond": 0.7280730336241769,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3335.214665998137,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892321368375886,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2012.1573750002426,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1387.5751249997847,
      "inferenceMs": 1384.5172920009645,
      "imagesPerSecond": 0.7222733914393779,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3409.2003749992728,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.741491350405051,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2012.1573750002426,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1807.0420409985672,
      "inferenceMs": 1803.2633749990055,
      "imagesPerSecond": 0.5545501638109583,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3829.5440419988154,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2012.1573750002426,
      "device": "cpu",
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "placement": {
        "cpu/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1381.2404999989667,
      "inferenceMs": 1378.146791999825,
      "imagesPerSecond": 0.7256121088152756,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3403.02658300061,
      "grade": "B"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 419.19679199963866,
      "inferenceMs": 405.5973750000703,
      "imagesPerSecond": 2.465499191161744,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 2024947712,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2698.241666001195,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7330345384033236,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 656.7188749995694,
      "inferenceMs": 635.0384999987,
      "imagesPerSecond": 1.574707675207168,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 3848863744,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2938.4909159998642,
      "grade": "B"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1902.2468750008557,
      "inferenceMs": 1892.8192499988654,
      "imagesPerSecond": 0.5283124630101894,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 4184.291124000083,
      "grade": "B"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 422.4960840001586,
      "inferenceMs": 408.75929199864913,
      "imagesPerSecond": 2.446427566479161,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2700.7911250002508,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 411.5239159982593,
      "inferenceMs": 400.5277500000375,
      "imagesPerSecond": 2.496705908641552,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2689.7777910016885,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 419.08845900070446,
      "inferenceMs": 405.02866599854315,
      "imagesPerSecond": 2.4689610488053626,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2697.507790999225,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 436.4938329999859,
      "inferenceMs": 423.5700419994828,
      "imagesPerSecond": 2.360884625549654,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2715.6959580006514,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2268.70745800079,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 418.7168330008717,
      "inferenceMs": 406.303208999816,
      "imagesPerSecond": 2.4612161012994926,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2697.2306660009053,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2411.9087500002934,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 424.7263750003185,
      "inferenceMs": 412.3482499999227,
      "imagesPerSecond": 2.425134579812543,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 2024947712,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2846.669500000644,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7330345384033236,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2411.9087500002934,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 763.87854099994,
      "inferenceMs": 739.9045420006587,
      "imagesPerSecond": 1.3515256945119682,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 3848863744,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 3186.9231250002485,
      "grade": "B"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2411.9087500002934,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 545.7645840015175,
      "inferenceMs": 535.9435840000515,
      "imagesPerSecond": 1.8658680313633607,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2967.2767920001206,
      "grade": "B"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9916942843547475,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2318.677833998663,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 425.77416700078174,
      "inferenceMs": 413.88670800006366,
      "imagesPerSecond": 2.4161201137192503,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 2024947712,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2754.4418339984986,
      "grade": "A"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7330345384033236,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2318.677833998663,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 691.0793340011878,
      "inferenceMs": 667.6701249998587,
      "imagesPerSecond": 1.4977456120269148,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 3848863744,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 3020.4473759986286,
      "grade": "B"
    }
  },
  {
    "precision": "bf16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2318.677833998663,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.bfloat16": 1184
      },
      "parameterDtype": "torch.bfloat16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 497.1585419989424,
      "inferenceMs": 489.2204589996254,
      "imagesPerSecond": 2.044068234686738,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 3850092544,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "bf16",
      "totalMs": 2826.422541998909,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 432.20191600084945,
      "inferenceMs": 419.25991700009035,
      "imagesPerSecond": 2.3851552687298376,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 2023391232,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2968.9509170002566,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7406320483415526,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 657.1393750000425,
      "inferenceMs": 639.8299579996092,
      "imagesPerSecond": 1.562915251931062,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 3845406720,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 3194.293625001592,
      "grade": "B"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 1982.0753750009317,
      "inferenceMs": 1971.720749999804,
      "imagesPerSecond": 0.507171210730576,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 4518.247166999572,
      "grade": "B"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 419.23016699911386,
      "inferenceMs": 405.134624999846,
      "imagesPerSecond": 2.4683153161751608,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2956.299958001182,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 434.42283299918927,
      "inferenceMs": 422.6622089990997,
      "imagesPerSecond": 2.365955551995258,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2971.06929200163,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 420.101457999408,
      "inferenceMs": 408.6728750007751,
      "imagesPerSecond": 2.4469448822560182,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2956.9343750008557,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 424.3117499991058,
      "inferenceMs": 412.34491700015496,
      "imagesPerSecond": 2.425154182268298,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2961.455874999956,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2527.0927920009854,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 412.10754200074007,
      "inferenceMs": 400.1565409998875,
      "imagesPerSecond": 2.4990220014928637,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2950.027417000456,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2277.7996670010907,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 404.65245799896365,
      "inferenceMs": 392.4804580001364,
      "imagesPerSecond": 2.5478975567228175,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 2023391232,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2691.864042000816,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7406320483415526,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2277.7996670010907,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 660.1995410001109,
      "inferenceMs": 641.5986250012793,
      "imagesPerSecond": 1.5586068314875303,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 3845406720,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2948.2298330003687,
      "grade": "B"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2277.7996670010907,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 483.261916999254,
      "inferenceMs": 476.5821250002773,
      "imagesPerSecond": 2.0982742481149876,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2770.5714590010757,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9888715423577618,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2337.6282079989323,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 403.0257089998486,
      "inferenceMs": 391.1803749997489,
      "imagesPerSecond": 2.556365461842614,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 2023391232,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2750.076498999988,
      "grade": "A"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7406320483415526,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2337.6282079989323,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 656.2660409999808,
      "inferenceMs": 637.2367909989407,
      "imagesPerSecond": 1.5692753684739185,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 3845406720,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 3004.2428739998286,
      "grade": "B"
    }
  },
  {
    "precision": "fp16",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2337.6282079989323,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float16": 1184
      },
      "parameterDtype": "torch.float16",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 473.3075000003737,
      "inferenceMs": 466.11974999905215,
      "imagesPerSecond": 2.1453714415706124,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 3845603328,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp16",
      "totalMs": 2821.490249998533,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 494.38775000089663,
      "inferenceMs": 480.45641600037925,
      "imagesPerSecond": 2.0813542429605323,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 4271538176,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3042.3940410000796,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7415304107602233,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 761.7800419993728,
      "inferenceMs": 740.5684999994264,
      "imagesPerSecond": 1.3503139817596543,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 6093553664,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3310.643374999927,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 518.371666999883,
      "inferenceMs": 511.6668339996977,
      "imagesPerSecond": 1.9543967549801964,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3066.54300000082,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 486.68937500042375,
      "inferenceMs": 473.7124160001258,
      "imagesPerSecond": 2.110985412718704,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3034.6292079993873,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 476.9587920000049,
      "inferenceMs": 464.93912500045553,
      "imagesPerSecond": 2.150819206705868,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3026.8022079999355,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 471.44370900059585,
      "inferenceMs": 458.32845800032374,
      "imagesPerSecond": 2.1818413902618583,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3019.763917000091,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 478.09679199963284,
      "inferenceMs": 462.25337500072783,
      "imagesPerSecond": 2.163315735657799,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3026.9696670002304,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2538.1997080003202,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 469.9119590004557,
      "inferenceMs": 457.75529200000165,
      "imagesPerSecond": 2.184573324386595,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3021.4695000013307,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2390.112417000637,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 460.03974999985076,
      "inferenceMs": 447.42245800080127,
      "imagesPerSecond": 2.2350241524939483,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 4271538176,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 2859.991084000285,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7415304107602233,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2390.112417000637,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 702.2935830009374,
      "inferenceMs": 686.9086249989778,
      "imagesPerSecond": 1.4557977052646387,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 6093553664,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3104.501459001767,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2390.112417000637,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 521.8218749996595,
      "inferenceMs": 515.2341670000169,
      "imagesPerSecond": 1.9408650746563691,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 2922.4817500016798,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.9892310711262283,
      "threshold": 0.7,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2372.438999998849,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 468.8425419990381,
      "inferenceMs": 456.56058300119184,
      "imagesPerSecond": 2.1902898262187245,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "dog. airplane.",
      "driverAllocatedMemoryBytes": 4271538176,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 2851.399457998923,
      "grade": "A"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "best matched-label box IoU + confidence + absent-label rejection",
      "score": 0.7415304107602233,
      "threshold": 0.4,
      "absentLabelPassed": true,
      "validBoxes": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2372.438999998849,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 703.8560830005736,
      "inferenceMs": 689.1425829999207,
      "imagesPerSecond": 1.45107852085831,
      "processedShape": [
        1,
        3,
        800,
        1199
      ],
      "inputTokens": 6,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "person. dog.",
      "driverAllocatedMemoryBytes": 6093553664,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 3088.1940419985767,
      "grade": "B"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "passed": true,
      "metric": "absent-object false positives",
      "score": 1,
      "expectedCount": 0,
      "actualCount": 0
    },
    "runtime": {
      "engine": "transformers",
      "version": "5.22.1",
      "torchVersion": "2.10.0",
      "loadMs": 2372.438999998849,
      "device": "mps",
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "placement": {
        "mps:0/torch.float32": 1184
      },
      "parameterDtype": "torch.float32",
      "customKernels": false,
      "accelerationDetails": "Native PyTorch deformable attention; parameters/input tensors on selected device. Image decoding, processor, tokenization and returned-box conversion use CPU.",
      "durationMs": 511.73662500150385,
      "inferenceMs": 504.83100000019476,
      "imagesPerSecond": 1.980860921773057,
      "processedShape": [
        1,
        3,
        800,
        970
      ],
      "inputTokens": 4,
      "threshold": 0.4,
      "textThreshold": 0.25,
      "normalizeLabels": true,
      "normalizedPrompt": "airplane.",
      "driverAllocatedMemoryBytes": 6093750272,
      "memoryMeasurementScope": "Post-inference driver allocation, not peak",
      "artifactRevision": "12bdfa3120f3e7ec7b434d90674b3396eccf88eb",
      "precision": "fp32",
      "totalMs": 2894.6356659980665,
      "grade": "B"
    }
  }
]

Something not quite right?

Tell us what happened with Grounding DINO Desktop. 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.