---
dataset_info:
  features:
  - name: observation.state
    sequence: float32
  - name: action
    sequence: float32
  - name: episode_id
    dtype: int64
  - name: frame_id
    dtype: int64
  - name: timestamp
    dtype: float32
  - name: next.done
    dtype: bool
  - name: observation.images.top
    sequence:
      sequence:
        sequence: uint8
  - name: index
    dtype: int64
  - name: episode_data_id_from
    dtype: int64
  - name: episode_data_id_to
    dtype: int64
  splits:
  - name: train
    num_bytes: 23188503125
    num_examples: 25000
  download_size: 1818503312
  dataset_size: 23188503125
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---