README.md

---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: episode_idx
    dtype: int64
  - name: step_idx
    dtype: int64
  - name: timestamp
    dtype: float64
  - name: observation
    struct:
    - name: image
      struct:
      - name: bytes
        dtype: image
    - name: camera
      struct:
      - name: intrinsic
        struct:
        - name: fx
          dtype: float64
        - name: fy
          dtype: float64
        - name: cx
          dtype: float64
        - name: cy
          dtype: float64
      - name: distortion
        struct:
        - name: k1
          dtype: float64
        - name: k2
          dtype: float64
        - name: p1
          dtype: float64
        - name: p2
          dtype: float64
        - name: k3
          dtype: float64
      - name: extrinsic
        struct:
        - name: rotation
          list:
            list: float64
        - name: translation
          list: float64
      - name: depth_scale
        dtype: float64
    - name: depth
      struct:
      - name: bytes
        dtype: image
    - name: extras
      struct:
      - name: categories
        list:
        - name: supercategory
          dtype: string
        - name: id
          dtype: int64
        - name: name
          dtype: string
  - name: action
    struct:
    - name: objects
      list:
      - name: name
        dtype: string
      - name: bbox_2d
        struct:
        - name: x1
          dtype: float64
        - name: y1
          dtype: float64
        - name: x2
          dtype: float64
        - name: y2
          dtype: float64
      - name: pose
        struct:
        - name: x
          dtype: float64
        - name: y
          dtype: float64
        - name: z
          dtype: float64
        - name: roll
          dtype: float64
        - name: pitch
          dtype: float64
        - name: yaw
          dtype: float64
  - name: image_keys
    dtype: string
  splits:
  - name: train
    num_bytes: 38650322179.875
    num_examples: 49995
  download_size: 38534215356
  dataset_size: 38650322179.875
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---