metadata
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': blipdataset
- name: text
dtype: string
splits:
- name: train
num_bytes: 456995431
num_examples: 29048
download_size: 461550035
dataset_size: 456995431
configs:
- config_name: default
data_files:
- split: train
path: data/train-*