metadata
dataset_info:
features:
- name: input_img
dtype: image
- name: processed_img
dtype: image
- name: instruction
dtype: string
- name: auxiliary_prompt
dtype: string
- name: task_id
dtype: string
splits:
- name: train
num_bytes: 406403776859.64
num_examples: 1013320
download_size: 416440124474
dataset_size: 406403776859.64
configs:
- config_name: default
data_files:
- split: train
path: data/train-*