metadata
dataset_info:
features:
- name: image_id
dtype: int64
- name: input
sequence: string
- name: output
dtype: string
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 271628169
num_examples: 566747
download_size: 25661831
dataset_size: 271628169
configs:
- config_name: default
data_files:
- split: train
path: data/train-*