metadata
dataset_info:
features:
- name: lit_image
dtype: image
- name: delit_image
dtype: image
- name: normal_image
dtype: image
- name: prompts
dtype: string
- name: text
dtype: string
- name: SH_idx
sequence: float32
- name: rot_idx
sequence: float32
splits:
- name: train
num_bytes: 765049954.418
num_examples: 5999
download_size: 763339342
dataset_size: 765049954.418
configs:
- config_name: default
data_files:
- split: train
path: data/train-*