---
dataset_info:
  features:
  - name: episode
    dtype: int64
  - name: frame_image
    dtype: image
  - name: action
    dtype: int64
  - name: next_frame_image
    dtype: image
  - name: done
    dtype: bool
  splits:
  - name: train
    num_bytes: 65293185.75
    num_examples: 3549
  download_size: 64337496
  dataset_size: 65293185.75
configs:
- config_name: default
  data_files:
  - split: train
    path: data/episode_920_episode_939/train-*
---