{
  "kind": "offline_checkpoint_replay",
  "checkpoint_archive": "checkpoint_dp.zip",
  "checkpoint_member": "outputs/blah/checkpoints/latest.ckpt",
  "checkpoint_sha256": "07ba7644d855c0435a695139114f6f35e1895b501bfcc6f737fd21196fad8cb0",
  "state_key": "ema_model",
  "episode": 14,
  "seed": 42,
  "inference_steps": 16,
  "candidates": 6,
  "horizon": 16,
  "n_obs_steps": 2,
  "n_action_steps": 15,
  "action_order": [
    "x",
    "y",
    "z",
    "rx",
    "ry",
    "rz",
    "left_jaw",
    "right_jaw"
  ],
  "alignment": "Approximate: video elapsed time matched to replay-buffer episode elapsed time; original camera timestamps unavailable.",
  "interpretation": "Actual model denoising iterates on recorded observations; these newly sampled actions were not executed on the robot and are not the original deployment predictions.",
  "samples": [
    {
      "time": 5.0,
      "sample_index": 50
    },
    {
      "time": 12.0,
      "sample_index": 120
    },
    {
      "time": 20.0,
      "sample_index": 200
    }
  ],
  "torch_version": "1.12.1",
  "policy_config": {
    "_target_": "diffusion_policy.policy.diffusion_unet_image_policy.DiffusionUnetImagePolicy",
    "shape_meta": {
      "obs": {
        "camera_0": {
          "shape": [
            3,
            240,
            320
          ],
          "type": "rgb"
        },
        "camera_1": {
          "shape": [
            3,
            240,
            320
          ],
          "type": "rgb"
        },
        "camera_2": {
          "shape": [
            3,
            240,
            320
          ],
          "type": "rgb"
        },
        "robot_eef_pose": {
          "shape": [
            6
          ],
          "type": "low_dim"
        },
        "left_jaw": {
          "shape": [
            1
          ],
          "type": "low_dim"
        },
        "right_jaw": {
          "shape": [
            1
          ],
          "type": "low_dim"
        }
      },
      "action": {
        "shape": [
          8
        ]
      }
    },
    "noise_scheduler": {
      "_target_": "diffusers.schedulers.scheduling_ddim.DDIMScheduler",
      "num_train_timesteps": 100,
      "beta_start": 0.0001,
      "beta_end": 0.02,
      "beta_schedule": "squaredcos_cap_v2",
      "clip_sample": true,
      "set_alpha_to_one": true,
      "steps_offset": 0,
      "prediction_type": "epsilon"
    },
    "obs_encoder": {
      "_target_": "diffusion_policy.model.vision.multi_image_obs_encoder.MultiImageObsEncoder",
      "shape_meta": {
        "obs": {
          "camera_0": {
            "shape": [
              3,
              240,
              320
            ],
            "type": "rgb"
          },
          "camera_1": {
            "shape": [
              3,
              240,
              320
            ],
            "type": "rgb"
          },
          "camera_2": {
            "shape": [
              3,
              240,
              320
            ],
            "type": "rgb"
          },
          "robot_eef_pose": {
            "shape": [
              6
            ],
            "type": "low_dim"
          },
          "left_jaw": {
            "shape": [
              1
            ],
            "type": "low_dim"
          },
          "right_jaw": {
            "shape": [
              1
            ],
            "type": "low_dim"
          }
        },
        "action": {
          "shape": [
            8
          ]
        }
      },
      "rgb_model": {
        "_target_": "diffusion_policy.model.vision.model_getter.get_resnet",
        "name": "resnet18",
        "weights": null
      },
      "resize_shape": [
        240,
        320
      ],
      "crop_shape": [
        216,
        288
      ],
      "random_crop": true,
      "use_group_norm": true,
      "share_rgb_model": false,
      "imagenet_norm": true
    },
    "horizon": 16,
    "n_action_steps": 6,
    "n_obs_steps": 2,
    "num_inference_steps": 100,
    "obs_as_global_cond": true,
    "diffusion_step_embed_dim": 128,
    "down_dims": [
      512,
      1024,
      2048
    ],
    "kernel_size": 5,
    "n_groups": 8,
    "cond_predict_scale": true
  }
}