metadata
dataset_info:
features:
- name: observation.images.top
dtype: image
- name: observation.state
list: float32
- name: action
list: float32
- name: episode_index
dtype: int64
- name: frame_index
dtype: int64
- name: timestamp
dtype: float32
- name: next.done
dtype: bool
- name: index
dtype: int64
- name: task_index
dtype: int64
splits:
- name: train
num_bytes: 1274873219
num_examples: 45000
download_size: 1256168986
dataset_size: 1274873219
configs:
- config_name: default
data_files:
- split: train
path: data/train-*