metadata
dataset_info:
features:
- name: image
dtype: image
- name: file_name
dtype: string
- name: captions
dtype: string
- name: longest
dtype: string
splits:
- name: train
num_bytes: 999791763
num_examples: 6000
- name: validation
num_bytes: 160454698
num_examples: 1000
- name: test
num_bytes: 167432398
num_examples: 1000
download_size: 1291883854
dataset_size: 1327678859
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*