metadata
dataset_info:
features:
- name: document_id
dtype: string
- name: image
dtype:
image:
decode: false
- name: original
dtype: string
- name: blip2
dtype: string
- name: internvl2
dtype: string
- name: llavanext
dtype: string
- name: minicpm
dtype: string
- name: paligemma
dtype: string
- name: phi3v
dtype: string
splits:
- name: train
num_bytes: 10135503
num_examples: 292
download_size: 10071520
dataset_size: 10135503
configs:
- config_name: default
data_files:
- split: train
path: data/train-*