metadata
dataset_info:
features:
- name: id
dtype: string
- name: img
dtype: string
- name: words
sequence: string
- name: bbox
sequence:
sequence: int64
- name: query
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 213385271.5723992
num_examples: 44242
- name: test
num_bytes: 23710546.427600797
num_examples: 4916
download_size: 36716617
dataset_size: 237095818
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*