metadata
dataset_info:
features:
- name: question
dtype: string
- name: ground_truth
list:
- name: full_answer
dtype: string
- name: ui_elements
list:
- name: bounds
sequence: float64
- name: text
dtype: string
- name: vh_index
dtype: int64
- name: image
dtype: image
- name: ocr_texts
dtype: string
- name: output
dtype: string
- name: input
dtype: string
splits:
- name: train
num_bytes: 5477115407.5
num_examples: 68980
download_size: 2204564566
dataset_size: 5477115407.5
configs:
- config_name: default
data_files:
- split: train
path: data/train-*