metadata
dataset_info:
features:
- name: image_url
dtype: string
- name: actor
dtype: string
- name: caption
dtype: string
- name: vae_512px
sequence:
sequence:
sequence:
sequence: float16
- name: ratio
dtype: float64
- name: t5_prompt_embeds
sequence:
sequence:
sequence: float16
- name: vae_1024px
sequence:
sequence:
sequence:
sequence: float16
- name: vae_512px-16ch
sequence:
sequence:
sequence:
sequence: float16
splits:
- name: train
num_bytes: 32790420473
num_examples: 28539
download_size: 31544508888
dataset_size: 32790420473
configs:
- config_name: default
data_files:
- split: train
path: data/train-*