metadata
dataset_info:
features:
- name: sample_id
dtype: string
- name: prompt_text
dtype: string
- name: target
dtype: string
- name: images
list: image
- name: num_pages
dtype: int32
- name: line_count
dtype: int32
- name: target_position
dtype: string
- name: split_id
dtype: string
splits:
- name: train
num_bytes: 201067194
num_examples: 700
download_size: 190362491
dataset_size: 201067194
configs:
- config_name: default
data_files:
- split: train
path: data/train-*