metadata
dataset_info:
features:
- name: screenshot
dtype: string
- name: cropped_screenshot
dtype: string
- name: expected_action
dtype: string
- name: expected_context
dtype: string
splits:
- name: train
num_bytes: 1558581
num_examples: 6
download_size: 1343051
dataset_size: 1558581
configs:
- config_name: default
data_files:
- split: train
path: data/train-*