metadata
dataset_info:
features:
- name: image
dtype: image
- name: prompt
dtype: string
- name: task1
dtype: string
- name: task2
dtype: string
- name: task3
dtype: string
- name: image_path
dtype: string
- name: original_prompt
dtype: string
- name: url
dtype: string
splits:
- name: train
num_bytes: 3105581715584.25
num_examples: 2080814
download_size: 3103020055730
dataset_size: 3105581715584.25
configs:
- config_name: default
data_files:
- split: train
path: data/train-*