metadata
dataset_info:
features:
- name: image
dtype: image
- name: input_word
dtype: string
- name: ground_truth
dtype: string
- name: task_type
dtype: string
splits:
- name: train
num_bytes: 3705540
num_examples: 425
- name: validation
num_bytes: 430260
num_examples: 47
- name: test
num_bytes: 474479
num_examples: 53
download_size: 4534569
dataset_size: 4610279
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*