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.

Table Transformer

Locate tables in document images. Returns table boxes and scores; text recognition is separate.

Desktop inferencePrivate · Runs on your computer
About this model

Locate tables in document images. Returns table boxes and scores; text recognition is separate.

Created by
Microsoft
License
MIT
Model released
Mar 4, 2022
Source repository created
Oct 14, 2022
Model weights · FP32
115.3 MB

Best for

  • Locate tables on document pages
  • Crop table regions before a separate OCR or structure-recognition pipeline

Keep in mind

  • Returns table locations, not text, rows, columns or cells. OCR and structure recognition are separate models.
  • Misses and false positives are possible on complex layouts, scans, small or rotated tables. Review all detections before downstream use.
  • Measured on four authored documents only; no broad accuracy claim. Threshold 0.9 rejects repeated-text false positives seen at 0.5 but can miss low-confidence tables.
  • PNG/JPEG still images only, maximum 20 MB, 20 megapixels and 16–10000 pixels per edge. EXIF orientation is applied and alpha is composited onto white.
  • No incremental model output: the runtime returns the complete detection query set after one forward pass.

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/microsoft/table-transformer-detection
Release verification
Verified
Release evidence
Original detection weights released March 4, 2022. Explicit user-requested older-model exception; 2023 safetensors conversion is not a new original release.
Source updated
Sep 6, 2023
Native model files
https://huggingface.co/microsoft/table-transformer-detection
Files updated
Sep 6, 2023
Artifact revision
2357cbe2b5a5d1c03e54f32764f06058933b65ab

Desktop app

Sample result

Table Transformer Desktop

Run Table Transformer 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 / FP32Tested8.0 GiBRAM estimate
macOS Apple SiliconMPS / FP32Tested8.0 GiBRAM estimate
Windows Intel / AMDCPU / FP32Unverified8.0 GiBRAM estimate
macOS Intel / AMDCPU / FP32Unverified8.0 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.

Table Transformer Desktop

23 configurations tested
Passed23
Failed0
TestedNov 4, 2026
Runtimetransformers 4.57.6

Quality criteria

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

PrecisionDeviceResultPlatformInference / load
FP32mpsAcceleratedPassedVery fastmacos · aarch6437 ms / 776 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6438 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6438 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6439 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6439 ms / 818 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6440 ms / 805 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6440 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6440 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6441 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6447 ms / 0 ms
FP32mpsAcceleratedPassedVery fastmacos · aarch6452 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6473 ms / 380 ms
FP32cpuPassedVery fastmacos · aarch6474 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6475 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6475 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6475 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6476 ms / 390 ms
FP32cpuPassedVery fastmacos · aarch6476 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6477 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6478 ms / 0 ms
FP32cpuPassedVery fastmacos · aarch6486 ms / 0 ms
FP32cpuPassedFastmacos · aarch64154 ms / 399 ms
FP32cpuPassedFastmacos · aarch64176 ms / 498 ms
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-04T00:49:02.183781+00:00
Environment
{ "os": "macos", "arch": "aarch64", "hardware": "Apple M3 Max", "osVersion": "27.2", "runtime": "Transformers 4.57.6 / Torch 2.6.0", "interpreter": "App-owned CPython 3.13.18 model-private managed environment" }
Runtime
transformers 4.57.6
Artifact revision
2357cbe2b5a5d1c03e54f32764f06058933b65ab
Fixture
{
  "file": "model-samples/desktop/table-transformer/ruled.png",
  "size": 36857,
  "sha1": "2c6e2ce9f13c8b5aad179274e232c33f1238d186",
  "sha256": "d74a71ec06b1bec20bb3f5df8c43e45c0d9bc364a91c65cedcaf1e4f4af2acc3",
  "url": "https://assets.localai.tools/samples/v1/files/d74a71ec06b1bec20bb3f5df8c43e45c0d9bc364a91c65cedcaf1e4f4af2acc3/ruled.png"
}
Rerun when
Artifact/runtime revision; Preprocessing, threshold, detection or presentation changes; Fixture or quality gate changes; Relevant OS/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 497.8513749956619,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 176.44574999576434,
      "durationMs": 186.96341599570587,
      "imagesPerSecond": 5.667464362411707,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 806699008,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 1,
      "totalMs": 684.8147909913678,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 399.1403330001049,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 153.6128750012722,
      "durationMs": 163.56174999964423,
      "imagesPerSecond": 6.5098709987149075,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 810942464,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 562.7020829997491,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.810193116852421
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 78.26862498768605,
      "durationMs": 88.28758299932815,
      "imagesPerSecond": 12.776511662972606,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 811286528,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 88.28758299932815,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8124264513165044,
        0.8158462175511468
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 86.43716698861681,
      "durationMs": 96.28445899579674,
      "imagesPerSecond": 11.569097355211714,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 813842432,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 96.28445899579674,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 73.89775000046939,
      "durationMs": 83.9235839957837,
      "imagesPerSecond": 13.532211738431117,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 814088192,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 83.9235839957837,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 74.60875000106171,
      "durationMs": 84.34012497309595,
      "imagesPerSecond": 13.403253639630334,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 814333952,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 84.34012497309595,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 75.79408399760723,
      "durationMs": 85.62979201087728,
      "imagesPerSecond": 13.193641868296334,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 814366720,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 85.62979201087728,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 77.4452920013573,
      "durationMs": 87.6647499972023,
      "imagesPerSecond": 12.912340752520812,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 814366720,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 87.6647499972023,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 75.39983300375752,
      "durationMs": 85.1507089973893,
      "imagesPerSecond": 13.262628843623107,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 817070080,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 85.1507089973893,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 75.21141701727174,
      "durationMs": 85.007207992021,
      "imagesPerSecond": 13.29585373681176,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 817119232,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 85.007207992021,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 389.9027500010561,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 75.5135840154253,
      "durationMs": 85.71866698912345,
      "imagesPerSecond": 13.242650485185926,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 803700736,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 475.62141699017957,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 379.5587499917019,
      "placement": {
        "cpu": 287
      },
      "provider": "PyTorch CPU",
      "acceleration": "cpu",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 73.44954099971801,
      "durationMs": 83.43983301892877,
      "imagesPerSecond": 13.6147889610888,
      "visualizationMs": 0,
      "predictionDevice": "cpu",
      "boxDevice": "cpu",
      "inputDevice": "cpu",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 804864000,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": null,
      "device": "cpu",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 462.9985830106307,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 775.8271250058897,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 37.035916990134865,
      "durationMs": 84.96333300718106,
      "imagesPerSecond": 27.000816538884852,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 604897280,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277231104,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 860.7904580130707,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.810193084165522
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 52.17016598908231,
      "durationMs": 68.4049999981653,
      "imagesPerSecond": 19.16804328759987,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 606519296,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277231104,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 68.4049999981653,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8124263131078637,
        0.8158460437395196
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 40.64858300262131,
      "durationMs": 61.66545799351297,
      "imagesPerSecond": 24.601103559637316,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 607961088,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277231104,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 61.66545799351297,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 39.438167004846036,
      "durationMs": 53.7741250009276,
      "imagesPerSecond": 25.356148014615467,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 608419840,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277214720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 53.7741250009276,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 37.86916600074619,
      "durationMs": 53.86620800709352,
      "imagesPerSecond": 26.406707768010936,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 609845248,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277214720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 53.86620800709352,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 38.46033301670104,
      "durationMs": 56.16966600064188,
      "imagesPerSecond": 26.000814906250532,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 609861632,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277214720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 56.16966600064188,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 39.74220898817293,
      "durationMs": 57.28687497321516,
      "imagesPerSecond": 25.162164496130412,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 610189312,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277214720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 57.28687497321516,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 40.26700000395067,
      "durationMs": 61.29370798589662,
      "imagesPerSecond": 24.834231502269553,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 612204544,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277214720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 61.29370798589662,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 0,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 47.062083991477266,
      "durationMs": 67.12016699020751,
      "imagesPerSecond": 21.248527799599685,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 612204544,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277214720,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 67.12016699020751,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 817.9516250093002,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 39.48354098247364,
      "durationMs": 91.91362501587719,
      "imagesPerSecond": 25.327009055340053,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 604454912,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277231104,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 909.8652500251774,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  },
  {
    "precision": "fp32",
    "device": "mps",
    "status": "pass",
    "quality": {
      "matchedIoU": [
        0.8123490601736304
      ],
      "recall": 1,
      "falsePositives": 0,
      "passed": true
    },
    "runtime": {
      "engine": "transformers",
      "version": "4.57.6",
      "torchVersion": "2.6.0",
      "loadMs": 805.2427920047194,
      "placement": {
        "mps:0": 287
      },
      "provider": "Metal/MPS",
      "acceleration": "partial",
      "parameterDtype": "torch.float32",
      "accelerationDetails": "Network parameters, inputs, logits and boxes on selected device; image decoding/resize/postprocessing/SVG on CPU.",
      "inferenceMs": 39.72291701938957,
      "durationMs": 94.16037501068786,
      "imagesPerSecond": 25.174384839660178,
      "visualizationMs": 0,
      "predictionDevice": "mps:0",
      "boxDevice": "mps:0",
      "inputDevice": "mps:0",
      "processedShape": [
        1,
        3,
        800,
        582
      ],
      "peakHelperRssBytes": 606994432,
      "memoryScope": "Process-lifetime peak RSS on macOS, not isolated model memory or GPU minimum",
      "driverAllocatedMemoryBytes": 1277231104,
      "device": "mps",
      "precision": "fp32",
      "cpuThreads": 4,
      "totalMs": 899.4031670154072,
      "grade": "fast",
      "fallback": false,
      "loadTimingScope": "Model construction/placement including one white 800x1100 image warmup after imports; excludes process/runtime/artifact setup",
      "totalTimingScope": "Model load plus full preprocessing/inference/postprocessing duration",
      "artifactRevision": "2357cbe2b5a5d1c03e54f32764f06058933b65ab"
    }
  }
]

Something not quite right?

Tell us what happened with Table Transformer. 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.