---
dataset_info:
  features:
  - name: episode_id
    dtype: int64
  - name: frames
    dtype: binary
  - name: actions
    dtype: int64
  - name: health
    dtype: int64
  - name: step_ids
    dtype: int64
  splits:
  - name: train
    num_bytes: 48664475554
    num_examples: 2478488
  download_size: 48429693358
  dataset_size: 48664475554
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---