{
  "provenance": {
    "architecture": "Original SAM ViT-B/16: 12 blocks, encoder width 768, prompt width 256",
    "parameterCount": 93735472,
    "sourceRevision": "dca509fe793f601edb92606367a655c15ac00fdf",
    "nativeSources": {
      "repository": "facebookresearch/segment-anything",
      "revision": "dca509fe793f601edb92606367a655c15ac00fdf",
      "files": {
        "LICENSE": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/LICENSE",
          "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4",
          "bytes": 11357
        },
        "segment_anything/build_sam.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/build_sam.py",
          "sha256": "77bc4faf728118625ddd184dc4e6375bd1b5c636fc66fb429f9fc5276f331cc5",
          "bytes": 2941
        },
        "segment_anything/modeling/__init__.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/__init__.py",
          "sha256": "47f6363b86c0bfaa7a6abf723975136d17fae4a5f002d2814a5cce8834d18980",
          "bytes": 385
        },
        "segment_anything/modeling/common.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/common.py",
          "sha256": "59a53bfa5d0a4d446df7f2bb57d2e95808b130601592b934600fb663d256ce17",
          "bytes": 1479
        },
        "segment_anything/modeling/image_encoder.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/image_encoder.py",
          "sha256": "1665ada2902722c9a8cd566c840e324d623ef2b12f7d9ae9f38b24aed0e39886",
          "bytes": 14420
        },
        "segment_anything/modeling/mask_decoder.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/mask_decoder.py",
          "sha256": "6a2455e990515c67a061187c92c236b879b189f21ef6ebfce1627cae9a6b2f93",
          "bytes": 6615
        },
        "segment_anything/modeling/prompt_encoder.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/prompt_encoder.py",
          "sha256": "bd8f2f5903f0647f55c21c608bcbb334eec2901ebe76aa41061b1c513df4a06f",
          "bytes": 8594
        },
        "segment_anything/modeling/sam.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/sam.py",
          "sha256": "2d501173b7345f6829f745a2ade240c98026f6b3237bf6cf8e9473405e86c54e",
          "bytes": 7226
        },
        "segment_anything/modeling/transformer.py": {
          "url": "https://raw.githubusercontent.com/facebookresearch/segment-anything/dca509fe793f601edb92606367a655c15ac00fdf/segment_anything/modeling/transformer.py",
          "sha256": "566a8f6db8b608398de939548bd0f485b21752fda2bdda36900e6444ef938d11",
          "bytes": 8397
        }
      }
    },
    "checkpointUrl": "https://dl.fbaipublicfiles.com/segment_anything/sam_vit_b_01ec64.pth",
    "checkpointSha256": "ec2df62732614e57411cdcf32a23ffdf28910380d03139ee0f4fcbe91eb8c912",
    "strictStateLoad": true,
    "missingKeys": [],
    "unexpectedKeys": [],
    "programSha256": "2b69ee67fd4a546e91efe7fb040a7665bad95086a666c85274f00f3f2a94e47b",
    "exportSha256": "46730d4bf8ad45fb7d9be8dce13d1e889565d10e32078fd69877d9c46d1fc59b",
    "command": "examples/sam_export.py --checkpoint .context/paper-series/weights/sam_vit_b_01ec64.pth --photo-dir public/examples/sam-demo --output public/examples/sam-demo",
    "reproductionCommand": "python sam_export.py --checkpoint sam_vit_b_01ec64.pth --output sam-demo",
    "python": "3.13.13",
    "torch": "2.12.0+cpu",
    "numpy": "2.1.2",
    "pillow": "12.2.0",
    "device": "CPU",
    "threads": 2,
    "seed": 7,
    "adapter": "Namespace-only import bypasses optional package APIs; native sources unchanged",
    "scope": "Two real-photo point prompts; not original benchmark or training reproduction",
    "scoreMeaning": "Predicted IoU regression for ranking; not calibrated probability"
  },
  "preprocessing": {
    "color": "RGB",
    "pixelRange": "0..255 float32",
    "targetLongestSide": 1024,
    "resize": "Pillow BILINEAR, official half-up dimensions",
    "mean": [
      123.675,
      116.28,
      103.53
    ],
    "std": [
      58.395,
      57.12,
      57.375
    ],
    "padding": "Right and bottom zeros after normalization to 1024 square",
    "coordinates": "Original (x,y) scaled by independently rounded width/height, then +0.5",
    "postprocess": "Bilinear logits to 1024 square, crop resized H/W, bilinear to original H/W",
    "alignCorners": false,
    "maskLogitThreshold": 0.0
  },
  "photos": [
    {
      "id": "cat",
      "title": "Chelsea the cat",
      "originalImage": "source-cat.png",
      "overlay": "cat-overlay.png",
      "overlaySha256": "2aabb739f809394d82845728cff7e43b39743fe7f59d0848043bf62f6490aa0e",
      "originalSize": [
        300,
        451
      ],
      "resizedSize": [
        681,
        1024
      ],
      "embeddingShape": [
        1,
        256,
        64,
        64
      ],
      "maskShape": [
        1,
        3,
        300,
        451
      ],
      "pointCoords": [
        240.0,
        160.0
      ],
      "pointLabels": [
        1
      ],
      "bestIndex": 0,
      "masks": [
        {
          "index": 0,
          "file": "cat-mask-0.png",
          "sha256": "9755c20f48c4d85108cff7ff5f4127f7903a2409107e89b9ccc3015d9f0770a9",
          "areaPixels": 104248,
          "predictedIoU": 0.98836749792099
        },
        {
          "index": 1,
          "file": "cat-mask-1.png",
          "sha256": "41006e989360843a5129ae7abae057178bd79cdc6b72a51009d9f4b216e68a63",
          "areaPixels": 83388,
          "predictedIoU": 0.9552326798439026
        },
        {
          "index": 2,
          "file": "cat-mask-2.png",
          "sha256": "e478fe31e1577f4c69f6099ab3862ee4be79a07508133448e27e4d38be7d532c",
          "areaPixels": 62099,
          "predictedIoU": 0.847079873085022
        }
      ],
      "arrays": "cat-outputs.npz",
      "arraysSha256": "54ef50f5db64cf37c71e14c65e633804652bc25aa0affd05b00bad8e06113498",
      "nativeMaxAbsErrors": [
        0.0,
        0.0,
        0.0
      ],
      "encoderSeconds": 8.15914561401587,
      "decoderSeconds": 0.046988343994598836,
      "reusedEmbedding": true,
      "refinementPredictedIoU": 0.8995227813720703,
      "source": {
        "url": "https://raw.githubusercontent.com/scikit-image/scikit-image/v0.25.2/skimage/data/chelsea.png",
        "sha256": "596aa1e7cb875eb79f437e310381d26b338a81c2da23439704a73c4651e8c4bb",
        "author": "Stefan van der Walt",
        "license": "CC0",
        "licenseUrl": "https://github.com/scikit-image/scikit-image/blob/v0.25.2/skimage/data/_fetchers.py"
      }
    },
    {
      "id": "coffee",
      "title": "Cup of coffee",
      "originalImage": "source-coffee.png",
      "overlay": "coffee-overlay.png",
      "overlaySha256": "648bfca5a3ef257f7728bea8d4c9e3db4188b9060bbe38d97641bd711b87e433",
      "originalSize": [
        400,
        600
      ],
      "resizedSize": [
        683,
        1024
      ],
      "embeddingShape": [
        1,
        256,
        64,
        64
      ],
      "maskShape": [
        1,
        3,
        400,
        600
      ],
      "pointCoords": [
        285.0,
        215.0
      ],
      "pointLabels": [
        1
      ],
      "bestIndex": 1,
      "masks": [
        {
          "index": 0,
          "file": "coffee-mask-0.png",
          "sha256": "9fa21e1cdcea1a8c200c429f1ffc27adc001babdcaa7d51be39bb08090a82653",
          "areaPixels": 95107,
          "predictedIoU": 0.8355532884597778
        },
        {
          "index": 1,
          "file": "coffee-mask-1.png",
          "sha256": "057e88f84b5251ec52c2485364222805adef50d699a1c67b255e735e37562ad9",
          "areaPixels": 37486,
          "predictedIoU": 0.9427707195281982
        },
        {
          "index": 2,
          "file": "coffee-mask-2.png",
          "sha256": "22285fbc49c48213a058cad22c747b913a2def8333c5719a35c5418978c18074",
          "areaPixels": 13622,
          "predictedIoU": 0.8657687902450562
        }
      ],
      "arrays": "coffee-outputs.npz",
      "arraysSha256": "4aa70e500388cc35b911c5a09087fa690508cdea7b8ea9df4be1e6d977bfeca2",
      "nativeMaxAbsErrors": [
        0.0,
        0.0,
        0.0
      ],
      "encoderSeconds": 8.294838199974038,
      "decoderSeconds": 0.0474091520300135,
      "reusedEmbedding": true,
      "refinementPredictedIoU": 0.6634522676467896,
      "source": {
        "url": "https://raw.githubusercontent.com/scikit-image/scikit-image/v0.25.2/skimage/data/coffee.png",
        "sha256": "cc02f8ca188b167c775a7101b5d767d1e71792cf762c33d6fa15a4599b5a8de7",
        "author": "Rachel Michetti",
        "license": "CC0",
        "licenseUrl": "https://github.com/scikit-image/scikit-image/blob/v0.25.2/skimage/data/_fetchers.py"
      }
    }
  ]
}