metadata
dataset_info:
features:
- name: episode_id
dtype: string
- name: messages
dtype: string
- name: images
sequence: image
- name: task
dtype: string
- name: instruction
dtype: string
- name: app
dtype: string
splits:
- name: train
num_bytes: 38166914
num_examples: 10
download_size: 37975876
dataset_size: 38166914
configs:
- config_name: default
data_files:
- split: train
path: data/train-*