metadata
dataset_info:
features:
- name: path
dtype: image
- name: question
dtype: string
- name: reasoning_trace
dtype: string
splits:
- name: train
num_bytes: 2334553291
num_examples: 10000
download_size: 2274721765
dataset_size: 2334553291
configs:
- config_name: default
data_files:
- split: train
path: data/train-*