metadata
dataset_info:
features:
- name: chunk_id
dtype: string
- name: text
dtype: string
- name: metadata
struct:
- name: Header 1
dtype: string
- name: Header 2
dtype: string
- name: image_references
list: string
- name: images_base64
list: string
- name: start_index
dtype: int64
- name: source_filename
dtype: string
splits:
- name: train
num_bytes: 32830551
num_examples: 315
download_size: 32605711
dataset_size: 32830551
configs:
- config_name: default
data_files:
- split: train
path: data/train-*