metadata
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: 74042828
num_examples: 4008
download_size: 72966045
dataset_size: 74042828
configs:
- config_name: default
data_files:
- split: train
path: data/episode_80_episode_99/train-*