metadata
dataset_info:
features:
- name: questionId
dtype: int32
- name: question
dtype: string
- name: question_types
list: string
- name: image
dtype: image
- name: docId
dtype: int32
- name: ucsf_document_id
dtype: string
- name: ucsf_document_page_no
dtype: string
- name: answers
list: string
splits:
- name: train
num_bytes: 8962286469.747847
num_examples: 12536
download_size: 2712114308
dataset_size: 8962286469.747847
configs:
- config_name: default
data_files:
- split: train
path: data/train-*