metadata
dataset_info:
features:
- name: image
dtype: string
- name: caption
dtype: string
- name: sentids
dtype: string
- name: split
dtype: string
- name: img_id
dtype: int64
- name: filename
dtype: string
- name: text
dtype: string
splits:
- name: train
num_bytes: 6657115
num_examples: 4149
download_size: 3492382
dataset_size: 6657115
configs:
- config_name: default
data_files:
- split: train
path: data/train-*