metadata
dataset_info:
features:
- name: original_image
dtype: image
- name: background_image
dtype: image
- name: drums_image
dtype: image
- name: edit_instruction
dtype: string
- name: source
dtype: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 8346910909.884
num_examples: 66382
- name: test
num_bytes: 1048344739.984
num_examples: 8297
- name: valid
num_bytes: 1038064856.691
num_examples: 8297
download_size: 10553195707
dataset_size: 10433320506.559
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*