data_config.json

{
  "batch_size": 192,
  "num_workers": 8,
  "shuffle_buffer_size": 1000,
  "normalization_mapping": {
    "ACTION": "ACTIONCHUNK_QUANTILE",
    "ENV": "QUANTILE",
    "STATE": "QUANTILE",
    "VISUAL": "IDENTITY"
  },
  "features": {
    "action": {
      "type": "ACTION",
      "shape": "(20,)"
    },
    "observation.image.0": {
      "type": "VISUAL",
      "shape": "(3, 224, 224)"
    },
    "observation.image.1": {
      "type": "VISUAL",
      "shape": "(3, 224, 224)"
    },
    "observation.image.2": {
      "type": "VISUAL",
      "shape": "(3, 224, 224)"
    },
    "observation.state": {
      "type": "STATE",
      "shape": "(20,)"
    }
  },
  "clip_values": {
    "ENV": [
      -1.05,
      1.05
    ],
    "STATE": [
      -1.05,
      1.05
    ]
  },
  "chunk_size": null,
  "action_time_horizon_s": null,
  "min_action_chunk_size": 8,
  "observation_mapping": {
    "action.ee_6d_pos": "action",
    "observation.ee_6d_pos": "observation.state",
    "observation.images.agentview_image": "observation.image.0",
    "observation.images.left_wrist_image": "observation.image.1",
    "observation.images.right_wrist_image": "observation.image.2",
    "observation.language_instruction": "task"
  },
  "observation_whitelist": null,
  "observation_blacklist": null,
  "transform_mapping": {
    "action": [
      {
        "type": "convert_to_6d_actions",
        "action_key": "action",
        "action_type": "EE_6D_POS",
        "input_type": "Dict",
        "post_norm": false
      },
      {
        "type": "delta_actions",
        "state_key": "observation.state",
        "action_key": "action",
        "action_type": "EE_6D_POS",
        "input_type": "Dict",
        "post_norm": false,
        "relative_to_state": true,
        "use_absolute_grippers": false
      }
    ],
    "observation.image.0": [
      {
        "type": "resize_with_padding",
        "height": 224,
        "width": 224,
        "mode": "bilinear",
        "align_corners": false,
        "input_type": "Tensor"
      },
      {
        "type": "color_jitter",
        "brightness": 0.2,
        "contrast": [
          0.8,
          1.2
        ],
        "saturation": [
          0.8,
          1.2
        ],
        "hue": 0.05,
        "input_type": "Tensor"
      }
    ],
    "observation.image.1": [
      {
        "type": "resize_with_padding",
        "height": 224,
        "width": 224,
        "mode": "bilinear",
        "align_corners": false,
        "input_type": "Tensor"
      },
      {
        "type": "color_jitter",
        "brightness": 0.2,
        "contrast": [
          0.8,
          1.2
        ],
        "saturation": [
          0.8,
          1.2
        ],
        "hue": 0.05,
        "input_type": "Tensor"
      }
    ],
    "observation.image.2": [
      {
        "type": "resize_with_padding",
        "height": 224,
        "width": 224,
        "mode": "bilinear",
        "align_corners": false,
        "input_type": "Tensor"
      },
      {
        "type": "color_jitter",
        "brightness": 0.2,
        "contrast": [
          0.8,
          1.2
        ],
        "saturation": [
          0.8,
          1.2
        ],
        "hue": 0.05,
        "input_type": "Tensor"
      }
    ],
    "observation.state": [
      {
        "type": "convert_to_6d_actions",
        "action_key": "observation.state",
        "action_type": "EE_6D_POS",
        "input_type": "Dict",
        "post_norm": false
      }
    ]
  },
  "language_instruction": null,
  "training_mode": "ROBOT_FLOWMATCH",
  "action_type": "EE_6D_POS",
  "serialize_stats": true,
  "video_decoder_cache_size": null,
  "repo_id": "tabletopsim_combined",
  "video_root_dir": null,
  "streaming": false,
  "tolerance_s": 0.1,
  "use_imagenet_stats": false,
  "split_name": "train",
  "prefetch_factor": 2,
  "max_episodes": null
}