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: 66795281.5
num_examples: 3634
download_size: 65772492
dataset_size: 66795281.5
configs:
- config_name: default
data_files:
- split: train
path: data/episode_1680_episode_1699/train-*