metadata
dataset_info:
features:
- name: image
dtype: image
- name: th_sentences_raw
dtype: string
splits:
- name: train
num_bytes: 50380215.75
num_examples: 1002
- name: validation
num_bytes: 13377387
num_examples: 250
- name: test
num_bytes: 14312700
num_examples: 250
download_size: 16049522
dataset_size: 78070302.75
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*