metadata
dataset_info:
features:
- name: action_chunks
sequence:
sequence: float32
- name: batch_idx
dtype: int64
- name: sample_idx
dtype: int64
- name: episode_idx
dtype: int64
- name: frame_idx
dtype: int64
- name: obs_observation_images_top
sequence:
sequence:
sequence: float32
- name: obs_observation_images_wrist
sequence:
sequence:
sequence: float32
- name: obs_action
sequence: float32
- name: obs_observation_state
sequence: float32
- name: obs_timestamp
dtype: float64
- name: obs_frame_index
dtype: int64
- name: obs_episode_index
dtype: int64
- name: obs_index
dtype: int64
- name: obs_task_index
dtype: int64
- name: obs_task
dtype: string
splits:
- name: train
num_bytes: 11817486400
num_examples: 1600
download_size: 2266780523
dataset_size: 11817486400
configs:
- config_name: default
data_files:
- split: train
path: data/train-*