metadata
dataset_info:
features:
- name: task_id
dtype: string
- name: domain
dtype: string
- name: first_frame
dtype: image
- name: final_frame
dtype: image
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1897644
num_examples: 75
download_size: 1705086
dataset_size: 1897644
configs:
- config_name: default
data_files:
- split: train
path: data/train-*