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.

SAM ViT Base — Point-Prompt Segmentation

Select an object in a still image by clicking points to include or exclude. Returns the highest-scoring object mask and candidate scores; it does not name the selected object.

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

Select an object in a still image by clicking points to include or exclude. Returns the highest-scoring object mask and candidate scores; it does not name the selected object.

Created by
Meta
License
Apache-2.0
Model released
Unknown
Source repository created
Unknown
Model weights · FP16
188.7 MB

Best for

  • Interactive foreground selection and image editing prototypes.
  • Generating candidate masks around user-selected image points.

Keep in mind

  • Mask quality depends on image quality, prompt placement, object boundaries, and scene complexity.
  • Small, transparent, occluded, or visually ambiguous objects may be missed or merged.
  • Predicted IoU scores are confidence estimates, not guarantees of accurate segmentation.
  • Masks may reflect dataset biases and must not be used as the sole basis for safety-critical, medical, identity, or access 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
nielsr/sam-vit-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
Jan 11, 2024
Browser model files
Xenova/sam-vit-base
Files updated
Mar 18, 2026
Artifact revision
e9ff25bcbb89d0be0758fcdf28e91193bb7eb6e9

Input

IMAGE

Sample result

SAM ViT Base — Point-Prompt Segmentation

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.

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.

SAM ViT Base — Point-Prompt Segmentation

6 configurations tested
Passed5
Failed1
TestedSep 12, 2026
RuntimeTransformers.js 4.2.0

Quality criteria

Finite candidate scores, expected output dimensions, selected point inside the best mask, and a non-degenerate foreground ratio.

PrecisionDeviceResult
FP16GPU · WebGPUAcceleratedPassedModerate
FP32GPU · WebGPUAcceleratedPassedModerate
Q8GPU · WebGPUAcceleratedPassedVery slow
Q8CPU · WASMPassedVery slow
FP32CPU · WASMPassedVery slow
FP16CPU · WASMFailed—
Test setup & sample details
Quality gate
Finite candidate scores, expected output dimensions, selected point inside the best mask, and a non-degenerate foreground ratio.
Tested
2026-09-12T18:46:29.238Z
Environment
Isolated headless Chrome on macOS; Transformers.js 4.2.0
Runtime
Transformers.js 4.2.0
Artifact revision
e9ff25bcbb89d0be0758fcdf28e91193bb7eb6e9
Fixture
{
  "file": "slimsam-77-uniform.json",
  "url": "https://f005.backblazeb2.com/file/hubters-webai-public/assets/benchmarks/slimsam-77-uniform.json",
  "size": 404,
  "sha1": "f03c558b1d0dfd0dcb2db0dcaffbd9038aac6cf8"
}
Rerun when
artifact revision change; runtime version change; fixture or quality gate change; browser/backend change
Per-case quality & failure details
[
  {
    "precision": "q8",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 14410.80000000447,
      "inferenceMs": 22917.29999999702,
      "totalMs": 38542.39999999851,
      "megapixelsPerSecond": 0.0305,
      "grade": "very-slow"
    }
  },
  {
    "precision": "q8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 19022.80000000447,
      "inferenceMs": 17576.39999999851,
      "totalMs": 37910.59999999404,
      "megapixelsPerSecond": 0.0398,
      "grade": "very-slow"
    }
  },
  {
    "precision": "fp16",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": "Not embedded",
    "failure": "Error: Can't create a session. ERROR_CODE: 1, ERROR_MESSAGE: /mnt/vss/_work/1/s/onnxruntime/core/graph/graph_utils.cc:30 int onnxruntime::graph_utils::GetIndexFromName(const Node &, const std::string &, bool) itr != node_args.end() was false. Attempting to get index by a name which does not exist:InsertedPrecisionFreeCast_/mask_decoder/upscale_layer_norm/Constant_output_0for node: /mask_decoder/transformer/layers.0/layer_norm1/Mul/SimplifiedLayerNormFusion/\n"
  },
  {
    "precision": "fp16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 26480.80000000447,
      "inferenceMs": 994.6999999955297,
      "totalMs": 28721.5,
      "megapixelsPerSecond": 0.7031,
      "grade": "slow"
    }
  },
  {
    "precision": "fp32",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 42049.29999999702,
      "inferenceMs": 23240.39999999851,
      "totalMs": 66679.5,
      "megapixelsPerSecond": 0.0301,
      "grade": "very-slow"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 41666.5,
      "inferenceMs": 1280.6000000014901,
      "totalMs": 44233.10000000149,
      "megapixelsPerSecond": 0.5461,
      "grade": "slow"
    }
  }
]

SAM ViT Large — Point-Prompt Segmentation

6 configurations tested
Passed4
Failed2
TestedSep 12, 2026
RuntimeTransformers.js 4.2.0

Quality criteria

Finite candidate scores, expected output dimensions, selected point inside the best mask, and a non-degenerate foreground ratio.

PrecisionDeviceResult
FP16GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
Q8GPU · WebGPUAcceleratedPassedVery slow
Q8CPU · WASMPassedVery slow
FP16CPU · WASMFailed—
FP32CPU · WASMFailed—
Test setup & sample details
Quality gate
Finite candidate scores, expected output dimensions, selected point inside the best mask, and a non-degenerate foreground ratio.
Tested
2026-09-12T18:46:29.240Z
Environment
Isolated headless Chrome on macOS; Transformers.js 4.2.0
Runtime
Transformers.js 4.2.0
Artifact revision
bad3c600df2df287ffbd1592849efbad608399e6
Fixture
{
  "file": "slimsam-77-uniform.json",
  "url": "https://f005.backblazeb2.com/file/hubters-webai-public/assets/benchmarks/slimsam-77-uniform.json",
  "size": 404,
  "sha1": "f03c558b1d0dfd0dcb2db0dcaffbd9038aac6cf8"
}
Rerun when
artifact revision change; runtime version change; fixture or quality gate change; browser/backend change
Per-case quality & failure details
[
  {
    "precision": "q8",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 35894.39999999851,
      "inferenceMs": 67654.79999999702,
      "totalMs": 104807.20000000298,
      "megapixelsPerSecond": 0.0103,
      "grade": "very-slow"
    }
  },
  {
    "precision": "q8",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 36903.70000000298,
      "inferenceMs": 59227.79999999702,
      "totalMs": 97427.29999999702,
      "megapixelsPerSecond": 0.0118,
      "grade": "very-slow"
    }
  },
  {
    "precision": "fp16",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": "Not embedded",
    "failure": "Error: Can't create a session. ERROR_CODE: 1, ERROR_MESSAGE: /mnt/vss/_work/1/s/onnxruntime/core/graph/graph_utils.cc:30 int onnxruntime::graph_utils::GetIndexFromName(const Node &, const std::string &, bool) itr != node_args.end() was false. Attempting to get index by a name which does not exist:InsertedPrecisionFreeCast_/mask_decoder/upscale_layer_norm/Constant_output_0for node: /mask_decoder/transformer/layers.0/layer_norm1/Mul/SimplifiedLayerNormFusion/\n"
  },
  {
    "precision": "fp16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 75602.90000000596,
      "inferenceMs": 2328.39999999851,
      "totalMs": 79396.60000000149,
      "megapixelsPerSecond": 0.3004,
      "grade": "slow"
    }
  },
  {
    "precision": "fp32",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": "Not embedded",
    "failure": "Error: failed to call OrtRun(). ERROR_CODE: 6, ERROR_MESSAGE: std::bad_alloc"
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "grade": "segmentation-pass",
      "pointInside": true
    },
    "runtime": {
      "loadMs": 122515.80000000447,
      "inferenceMs": 2623.5,
      "totalMs": 126382.40000000596,
      "megapixelsPerSecond": 0.2666,
      "grade": "slow"
    }
  }
]

Something not quite right?

Tell us what happened with SAM 1. 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.