{
  "schema_version": 1,
  "interpretation": "Box/point-guided SAM 2.1 video masks for visualization; not LIBERO ground truth or SR-WM predictions.",
  "model": {
    "id": "facebook/sam2.1-hiera-small",
    "revision": "ee5bba1d82bb8749febdf90f45e84b687142ba03",
    "weights_file": "model.safetensors",
    "weights_sha256": "0a4067b11ce1e23d5229203f11c718a823060d15a4b23fa2372a7d4b77cbbc60",
    "license": "apache-2.0"
  },
  "roles": [
    "gripper",
    "target",
    "goal"
  ],
  "clips": [
    {
      "suite_key": "libero10",
      "source": "showcase-libero10-book-caddy.mp4",
      "source_sha256": "e009da8dea9df417d6f1416426a682d7bddc2f4a21d32c6eb3317d29c9380fc7",
      "output": "showcase-libero10-book-caddy-sam2-roles.mp4",
      "output_bytes": 2119741,
      "output_sha256": "6b6569e0edd7c017787b8391f9d7eae87f9e7d9653e21b00524d01af79b63889",
      "frame_count": 150,
      "fps": 10.0,
      "inference": {
        "device": "cpu",
        "dtype": "float32",
        "transformers_version": "4.57.1",
        "torch_version": "2.7.1+cpu"
      },
      "box_prompts_xyxy": [
        {
          "frame_idx": 0,
          "boxes": {
            "gripper": [
              190,
              55,
              405,
              185
            ],
            "target": [
              160,
              315,
              220,
              480
            ],
            "goal": [
              260,
              205,
              585,
              370
            ]
          }
        }
      ],
      "point_prompts_xy_labels": [
        {
          "frame_idx": 90,
          "roles": {
            "goal": {
              "points": [
                [
                  300,
                  340
                ],
                [
                  520,
                  340
                ],
                [
                  520,
                  255
                ],
                [
                  335,
                  245
                ],
                [
                  355,
                  190
                ]
              ],
              "labels": [
                1,
                1,
                1,
                0,
                0
              ]
            }
          }
        },
        {
          "frame_idx": 100,
          "roles": {
            "goal": {
              "points": [
                [
                  300,
                  340
                ],
                [
                  520,
                  340
                ],
                [
                  520,
                  255
                ],
                [
                  405,
                  210
                ],
                [
                  420,
                  145
                ]
              ],
              "labels": [
                1,
                1,
                1,
                0,
                0
              ]
            }
          }
        }
      ],
      "binary_mask_stream_sha256": "5bd8c103aade7424ddd374d990e35c1c6cef57ab422b7354df3ac016a90cc5ca",
      "mask_area_pixels": {
        "gripper": {
          "min": 2643,
          "max": 18872,
          "mean": 12162.39,
          "p05": 2909.65,
          "p95": 18295.1,
          "zero_area_frames": []
        },
        "target": {
          "min": 3199,
          "max": 11679,
          "mean": 6566.4,
          "p05": 3548.25,
          "p95": 10904.55,
          "zero_area_frames": []
        },
        "goal": {
          "min": 40212,
          "max": 49430,
          "mean": 46446.25,
          "p05": 41866.4,
          "p95": 49011.5,
          "zero_area_frames": []
        }
      }
    },
    {
      "suite_key": "goal",
      "source": "showcase-goal-stove.mp4",
      "source_sha256": "80bb68664fc3d07c9353358c54ec4af27586fd5beb7b753dabbe3d21b85341ba",
      "output": "showcase-goal-stove-sam2-roles.mp4",
      "output_bytes": 1052499,
      "output_sha256": "c79a8ae53b08d38b9715d0de09a6c8443e14b50b078ed1cf0fb4ced6c18ce8af",
      "frame_count": 90,
      "fps": 10.0,
      "inference": {
        "device": "cuda",
        "dtype": "bfloat16",
        "transformers_version": "4.57.1",
        "torch_version": "2.7.1+cu128"
      },
      "first_frame_boxes_xyxy": {
        "gripper": [
          215,
          0,
          430,
          215
        ],
        "target": [
          160,
          185,
          230,
          255
        ],
        "goal": [
          90,
          245,
          260,
          350
        ]
      },
      "binary_mask_stream_sha256": "6433661622132dd6190c38aad72991f4ad332380b0e0bac490840f675d185eb0",
      "mask_area_pixels": {
        "gripper": {
          "min": 33749,
          "max": 44883,
          "mean": 41445.28
        },
        "target": {
          "min": 906,
          "max": 2065,
          "mean": 1746.52
        },
        "goal": {
          "min": 11714,
          "max": 12068,
          "mean": 11900.38
        }
      }
    },
    {
      "suite_key": "object",
      "source": "showcase-object-milk.mp4",
      "source_sha256": "f471739156844da74f9a33f08b38e51d4dac8b0dd69c01e56aa646a7d975c111",
      "output": "showcase-object-milk-sam2-roles.mp4",
      "output_bytes": 1980062,
      "output_sha256": "fee255181da13c8b01e2f219e44ea0468ff777d91aafa1609d42e04fd25bf05b",
      "frame_count": 141,
      "fps": 10.0,
      "inference": {
        "device": "cuda",
        "dtype": "bfloat16",
        "transformers_version": "4.57.1",
        "torch_version": "2.7.1+cu128"
      },
      "first_frame_boxes_xyxy": {
        "gripper": [
          220,
          15,
          425,
          220
        ],
        "target": [
          445,
          235,
          515,
          345
        ],
        "goal": [
          25,
          250,
          190,
          420
        ]
      },
      "binary_mask_stream_sha256": "119ee19aa09cf38dfb2b2059c576ca46d40a3077d1a9273711554d688268ec71",
      "mask_area_pixels": {
        "gripper": {
          "min": 23946,
          "max": 48545,
          "mean": 39985.89
        },
        "target": {
          "min": 2255,
          "max": 4935,
          "mean": 3940.16
        },
        "goal": {
          "min": 16772,
          "max": 19606,
          "mean": 19097.23
        }
      }
    },
    {
      "suite_key": "spatial",
      "source": "showcase-spatial-bowl.mp4",
      "source_sha256": "1eb11fb58fafa1eab9d0f5d8fe043a771969626c64660b35660f34c8afcd3266",
      "output": "showcase-spatial-bowl-sam2-roles.mp4",
      "output_bytes": 1412683,
      "output_sha256": "3f81a933c37bcd7a650490adc95b6abd01099846fff90cc0691ca282d47b052f",
      "frame_count": 95,
      "fps": 10.0,
      "inference": {
        "device": "cuda",
        "dtype": "bfloat16",
        "transformers_version": "4.57.1",
        "torch_version": "2.7.1+cu128"
      },
      "first_frame_boxes_xyxy": {
        "gripper": [
          215,
          0,
          430,
          220
        ],
        "target": [
          95,
          345,
          200,
          430
        ],
        "goal": [
          70,
          425,
          230,
          515
        ]
      },
      "binary_mask_stream_sha256": "80cbffef1ac9e2daa68f9055c976acc7f2501a3a9a29aab8bf1ba836a39ca6d7",
      "mask_area_pixels": {
        "gripper": {
          "min": 35000,
          "max": 76326,
          "mean": 62425.11
        },
        "target": {
          "min": 1820,
          "max": 6126,
          "mean": 3952.56
        },
        "goal": {
          "min": 2731,
          "max": 8272,
          "mean": 7104.58
        }
      }
    }
  ]
}
