metadata
dataset_info:
features:
- name: image_id
dtype: string
- name: sentence_id
dtype: int64
- name: text
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 57705952912.112
num_examples: 360216
download_size: 19887197751
dataset_size: 57705952912.112
configs:
- config_name: default
data_files:
- split: train
path: data/train-*