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 Tiny — Open-Vocabulary Detection

Find objects in an image by describing what to look for, rather than choosing from a fixed class list. Returns matching boxes and confidence scores for your text labels.

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

Find objects in an image by describing what to look for, rather than choosing from a fixed class list. Returns matching boxes and confidence scores for your text labels.

Created by
IDEA Research
License
Apache-2.0
Model released
Unknown
Source repository created
Sep 25, 2023
Model weights · Q4
227.2 MB

Best for

  • Prototype prompt-conditioned object detection for images.
  • Locate uncommon object categories that are not present in a fixed detector label list.
  • Build browser image annotation, search, and accessibility experiments.

Keep in mind

  • Detection quality depends strongly on label wording; short lowercase noun phrases usually work better than abstract or ambiguous prompts.
  • Small, occluded, blurred, unusual, or densely packed objects may be missed or assigned inaccurate boxes.
  • Scores are not calibrated probabilities, and different label sets can change the returned detections.
  • The model may reflect biases in its training data and must not be used alone for surveillance, identity, safety-critical, or other high-impact 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
IDEA-Research/grounding-dino-tiny
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
Browser model files
onnx-community/grounding-dino-tiny-ONNX
Files updated
Jan 5, 2025
Artifact revision
ff690b0a8050566c290287545bd059350f3e9096

Input

IMAGE

Sample result

Grounding DINO Tiny — Open-Vocabulary Detection

Settings

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.
Advanced settings 1
Minimum detection score returned by post-processing. Lower values keep more weak predictions or foreground pixels, with more possible false positives. Higher values are stricter and may miss details. Automatic/null preserves soft alpha edges when offered.

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.

Grounding DINO Tiny — Open-Vocabulary Detection

16 configurations tested
Passed12
Failed4
TestedOct 23, 2026
RuntimeTransformers.js 4.2.0

Quality criteria

For prompts 'a person' and 'a dog', output must contain a finite 'a person' detection with score >= 0.3 and IoU >= 0.4 against the checked-in subject box, return no 'a dog' false positive above threshold, and contain only valid boxes.

PrecisionDeviceResult
Q4GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedVery slow
BNB4GPU · WebGPUAcceleratedPassedVery slow
INT8GPU · WebGPUAcceleratedPassedVery slow
Q8GPU · WebGPUAcceleratedPassedVery slow
UINT8GPU · WebGPUAcceleratedPassedVery slow
BNB4CPU · WASMPassedVery slow
Q4CPU · WASMPassedVery slow
INT8CPU · WASMPassedVery slow
FP32CPU · WASMPassedVery slow
Q8CPU · WASMPassedVery slow
UINT8CPU · WASMPassedVery slow
FP16CPU · WASMFailed—
FP16GPU · WebGPUFailed—
Q4F16CPU · WASMFailed—
Q4F16GPU · WebGPUFailed—
Test setup & sample details
Quality gate
For prompts 'a person' and 'a dog', output must contain a finite 'a person' detection with score >= 0.3 and IoU >= 0.4 against the checked-in subject box, return no 'a dog' false positive above threshold, and contain only valid boxes.
Tested
2026-10-23T13:29:12.152Z
Environment
Isolated headless Chrome on macOS; Transformers.js 4.2.0
Runtime
Transformers.js 4.2.0
Artifact revision
ff690b0a8050566c290287545bd059350f3e9096
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/"
}
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": {
      "valid": true,
      "bestIoU": 0.7976,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 60710.6,
      "inferenceMs": 13371.6,
      "totalMs": 74231,
      "grade": "very-slow",
      "imagesPerSecond": 0.0748
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.7976,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 68605.4,
      "inferenceMs": 5013.1,
      "totalMs": 73757.1,
      "grade": "very-slow",
      "imagesPerSecond": 0.1995
    }
  },
  {
    "precision": "fp16",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 36948,
      "grade": "failed"
    },
    "failure": "onnx-session-error"
  },
  {
    "precision": "fp16",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "valid": false,
      "bestIoU": 0,
      "falsePositiveCount": 0,
      "grade": "failed"
    },
    "runtime": {
      "loadMs": 35093,
      "inferenceMs": 4141.3,
      "totalMs": 39341.5,
      "grade": "slow",
      "imagesPerSecond": 0.2415
    },
    "failure": "quality-invalid-output"
  },
  {
    "precision": "int8",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.7914,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 15430.1,
      "inferenceMs": 13308.8,
      "totalMs": 28870.4,
      "grade": "very-slow",
      "imagesPerSecond": 0.0751
    }
  },
  {
    "precision": "int8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.7865,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 21755.5,
      "inferenceMs": 14438,
      "totalMs": 36323.8,
      "grade": "very-slow",
      "imagesPerSecond": 0.0693
    }
  },
  {
    "precision": "q8",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.8163,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 17387.4,
      "inferenceMs": 13376,
      "totalMs": 30906.4,
      "grade": "very-slow",
      "imagesPerSecond": 0.0748
    }
  },
  {
    "precision": "q8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.8163,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 16102.3,
      "inferenceMs": 14540.5,
      "totalMs": 31065.7,
      "grade": "very-slow",
      "imagesPerSecond": 0.0688
    }
  },
  {
    "precision": "uint8",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.8163,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 21907.2,
      "inferenceMs": 13459.6,
      "totalMs": 35490.7,
      "grade": "very-slow",
      "imagesPerSecond": 0.0743
    }
  },
  {
    "precision": "uint8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.8163,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 23051.7,
      "inferenceMs": 14656.1,
      "totalMs": 37966.6,
      "grade": "very-slow",
      "imagesPerSecond": 0.0682
    }
  },
  {
    "precision": "q4",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.783,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 19242,
      "inferenceMs": 13201.8,
      "totalMs": 32558.7,
      "grade": "very-slow",
      "imagesPerSecond": 0.0757
    }
  },
  {
    "precision": "q4",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.783,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 23187.2,
      "inferenceMs": 2605.2,
      "totalMs": 25899.6,
      "grade": "slow",
      "imagesPerSecond": 0.3838
    }
  },
  {
    "precision": "q4f16",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 20978.7,
      "grade": "failed"
    },
    "failure": "onnx-session-error"
  },
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "valid": false,
      "bestIoU": 0,
      "falsePositiveCount": 0,
      "grade": "failed"
    },
    "runtime": {
      "loadMs": 17378.5,
      "inferenceMs": 1608.8,
      "totalMs": 19184.6,
      "grade": "slow",
      "imagesPerSecond": 0.6216
    },
    "failure": "quality-invalid-output"
  },
  {
    "precision": "bnb4",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.7861,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 31170.3,
      "inferenceMs": 12612,
      "totalMs": 43886.9,
      "grade": "very-slow",
      "imagesPerSecond": 0.0793
    }
  },
  {
    "precision": "bnb4",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "valid": true,
      "bestIoU": 0.7861,
      "falsePositiveCount": 0,
      "grade": "grounded-detection-pass"
    },
    "runtime": {
      "loadMs": 28194.8,
      "inferenceMs": 13127.6,
      "totalMs": 41459.2,
      "grade": "very-slow",
      "imagesPerSecond": 0.0762
    }
  }
]

Something not quite right?

Tell us what happened with Grounding DINO. 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.