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.

ViT Face Emotion Detection — FER2013

Classify visible facial-expression patterns in a clear face image into seven categories. Returns ranked scores; these are predictions about appearance, not reliable evidence of someone's emotions.

Browser inferencePowered by WebAI · Private · Unlimited runs
About this model

Classify visible facial-expression patterns in a clear face image into seven categories. Returns ranked scores; these are predictions about appearance, not reliable evidence of someone's emotions.

Created by
Abhilash
License
Apache-2.0
Model released
Unknown
Source repository created
Jul 13, 2025
Model weights · Q4F16
49.8 MB

Best for

  • Research and educational exploration of facial-expression classification.
  • Benchmark browser image classification across ONNX variants.
  • Add non-consequential expression tags to user-reviewed media workflows.

Keep in mind

  • A facial expression is not reliable evidence of a person's internal emotional state, intent, truthfulness, or mental health.
  • The reported source accuracy is 71.55% on FER2013, and performance varies across demographic groups.
  • The model expects a clear, centered, front-facing face and can fail with multiple faces, profiles, occlusion, poor lighting, blur, or non-photographic images.
  • FER2013 contains low-resolution, noisy labels and only seven coarse classes.
  • Do not use this model for biometric identification, surveillance, employment, education, health, law enforcement, or other consequential decisions.

Processed on your deviceInference runs in this browser. Input/output sharing requires consent for each run. Basic run logs include device information and approximate connection location. Model and runtime files download from external hosts. Saved downloads may be cleared by your browser.

Sources & technical details
Original model
abhilash88/face-emotion-detection
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
Jul 13, 2025
Browser model files
onnx-community/face-emotion-detection-ONNX
Files updated
Aug 29, 2026
Artifact revision
905691b2f5f19e65b2dc909b01f0451a2e81d963

Input

IMAGE

Sample result

ViT Face Emotion Detection — FER2013

Settings

Number of ranked expression classes to return. Lower values show only the strongest classes. Higher values expose more alternatives, including lower-confidence labels, without improving the underlying prediction.
Adds a self-contained SVG presentation image alongside the unchanged native model fields. When enabled, draws model predictions or presents generated media and text. Adds CPU postprocessing time and image bytes. It does not change model predictions.

Output

Your result will appear here

Add input or try the sample, then run the model.

Benchmark evidence

Saved worker evidence is separate from benchmarks run on your device. It does not guarantee speed or quality on every browser.

ViT Face Emotion Detection — FER2013

16 configurations tested
Passed14
Failed2
TestedSep 15, 2026
RuntimeTransformers.js 4.2.0

Quality criteria

All seven unique labels must have finite probabilities summing to one; the checked-in fixture's top label and minimum probability margin must match the FP32 reference.

PrecisionDeviceResult
Q4F16GPU · WebGPUAcceleratedPassedVery fast
BNB4GPU · WebGPUAcceleratedPassedVery fast
Q4GPU · WebGPUAcceleratedPassedFast
FP32GPU · WebGPUAcceleratedPassedFast
FP16GPU · WebGPUAcceleratedPassedFast
UINT8GPU · WebGPUAcceleratedPassedModerate
INT8GPU · WebGPUAcceleratedPassedModerate
Q8GPU · WebGPUAcceleratedPassedModerate
FP32CPU · WASMPassedModerate
UINT8CPU · WASMPassedModerate
FP16CPU · WASMPassedModerate
BNB4CPU · WASMPassedModerate
Q4CPU · WASMPassedModerate
Q4F16CPU · WASMPassedModerate
INT8CPU · WASMFailed—
Q8CPU · WASMFailed—
Test setup & sample details
Quality gate
All seven unique labels must have finite probabilities summing to one; the checked-in fixture's top label and minimum probability margin must match the FP32 reference.
Tested
2026-09-15T14:44:28.380Z
Environment
Isolated headless Chrome on macOS; Transformers.js 4.2.0
Runtime
Transformers.js 4.2.0
Artifact revision
905691b2f5f19e65b2dc909b01f0451a2e81d963
Fixture
{
  "file": "modnet-portrait-pexels-5965592.jpg",
  "url": "https://f005.backblazeb2.com/file/hubters-webai-public/assets/benchmarks/modnet-portrait-pexels-5965592.jpg",
  "size": 78124,
  "sha1": "38212739a4a4f0be25d84863cf991d7fb6c6d818",
  "width": 1024,
  "height": 683,
  "source": "https://www.pexels.com/photo/woman-wearing-pink-turtleneck-sweater-5965592/",
  "license": "Pexels License",
  "licenseUrl": "https://www.pexels.com/license/",
  "preprocessing": "Pinned ViT image processor resizes to 224×224 and normalizes RGB channels with mean/std 0.5."
}
Rerun when
artifact revision change; runtime version change; fixture or quality-gate change; browser/backend behavior change
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.05456706881523132,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 40562,
      "inferenceMs": 837.7000000178814,
      "totalMs": 42687.59999999404,
      "grade": "slow",
      "imagesPerSecond": 1.1937447773411176
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.05456671118736267,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 39165.5,
      "inferenceMs": 171.5,
      "totalMs": 40559.90000000596,
      "grade": "good",
      "imagesPerSecond": 5.830903790087463
    }
  },
  {
    "precision": "fp16",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.05360826849937439,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 21296.69999998808,
      "inferenceMs": 847.3000000119209,
      "totalMs": 23458.20000001788,
      "grade": "slow",
      "imagesPerSecond": 1.1802195208142696
    }
  },
  {
    "precision": "fp16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.054856449365615845,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 21917.5,
      "inferenceMs": 246.7999999821186,
      "totalMs": 23512.19999998808,
      "grade": "good",
      "imagesPerSecond": 4.0518638576679615
    }
  },
  {
    "precision": "int8",
    "device": "wasm",
    "status": "fail",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.023201167583465576,
      "minimumMargin": 0.04,
      "grade": "failed"
    },
    "runtime": {
      "loadMs": 15993.59999999404,
      "inferenceMs": 871.0999999940395,
      "totalMs": 18209.5,
      "grade": "slow",
      "imagesPerSecond": 1.1479738262046177
    }
  },
  {
    "precision": "int8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.04910120368003845,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 14669.200000017881,
      "inferenceMs": 922.9000000059605,
      "totalMs": 16847.600000023842,
      "grade": "slow",
      "imagesPerSecond": 1.0835410120203073
    }
  },
  {
    "precision": "uint8",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.09218215942382812,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 15444.799999982119,
      "inferenceMs": 841.8000000119209,
      "totalMs": 17566,
      "grade": "slow",
      "imagesPerSecond": 1.1879306248346861
    }
  },
  {
    "precision": "uint8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.06777578592300415,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 14420.199999988079,
      "inferenceMs": 896.5999999940395,
      "totalMs": 16649.399999976158,
      "grade": "slow",
      "imagesPerSecond": 1.1153245594542136
    }
  },
  {
    "precision": "q8",
    "device": "wasm",
    "status": "fail",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.023201167583465576,
      "minimumMargin": 0.04,
      "grade": "failed"
    },
    "runtime": {
      "loadMs": 16021.09999999404,
      "inferenceMs": 868.3000000119209,
      "totalMs": 18109.40000000596,
      "grade": "slow",
      "imagesPerSecond": 1.1516756881104122
    }
  },
  {
    "precision": "q8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.04910120368003845,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 14693.90000000596,
      "inferenceMs": 923.9000000059605,
      "totalMs": 16869,
      "grade": "slow",
      "imagesPerSecond": 1.082368221662029
    }
  },
  {
    "precision": "q4",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.08844156563282013,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 10317.5,
      "inferenceMs": 937.7000000178814,
      "totalMs": 12520.399999976158,
      "grade": "slow",
      "imagesPerSecond": 1.0664391596256058
    }
  },
  {
    "precision": "q4",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.08844086527824402,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 11786,
      "inferenceMs": 105.09999999403954,
      "totalMs": 13147.09999999404,
      "grade": "good",
      "imagesPerSecond": 9.514747859721336
    }
  },
  {
    "precision": "q4f16",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.08765895664691925,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 11345.199999988079,
      "inferenceMs": 969.4000000059605,
      "totalMs": 13580.59999999404,
      "grade": "slow",
      "imagesPerSecond": 1.0315659170557576
    }
  },
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.08672153949737549,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 14555.699999988079,
      "inferenceMs": 96.30000001192093,
      "totalMs": 15941.199999988079,
      "grade": "excellent",
      "imagesPerSecond": 10.38421599040717
    }
  },
  {
    "precision": "bnb4",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.08844156563282013,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 9684.09999999404,
      "inferenceMs": 936.9000000059605,
      "totalMs": 11979,
      "grade": "slow",
      "imagesPerSecond": 1.067349770513009
    }
  },
  {
    "precision": "bnb4",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "finite": true,
      "uniqueLabels": true,
      "expectedTopLabel": "sad",
      "actualTopLabel": "sad",
      "margin": 0.08844086527824402,
      "minimumMargin": 0.04,
      "grade": "face-expression-pass"
    },
    "runtime": {
      "loadMs": 10375,
      "inferenceMs": 97,
      "totalMs": 11647.800000011921,
      "grade": "excellent",
      "imagesPerSecond": 10.309278350515465
    }
  }
]

Something not quite right?

Tell us what happened with ViT Face Emotion Detection. 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.