Views
No views yet
Status: This model is still in training. The current weights are a very early checkpoint and are shared mainly for testing, experimentation, and reproducible inference. Output quality, conditioning accuracy, and generation stability are expected to improve in future checkpoints.
artistgenrestyleANY_* token.1pip install -r requirements.txt
2python generate.py --model_path . --artist "pablo-picasso" --genre "portrait" --style "Cubism" --scheduler ddim --num_inference_steps 50 --guidance_scale 2.0 --num_images 16 --grid --seed 42python generate.py --model_path . --list_labels --label_limit 251--model_path Path to the cloned model repository.
2--artist Artist label. Omit for ANY_ARTIST.
3--genre Genre label. Omit for ANY_GENRE.
4--style Style label. Omit for ANY_STYLE.
5--num_images Number of images to generate.
6--batch_size Images per generation batch. 0 means all images at once.
7--scheduler ddpm, ddim, or dpm.
8--num_inference_steps Number of denoising steps.
9--guidance_scale Classifier-free guidance scale. 1.0 disables guidance.
10--output_dir Directory for generated images.
11--grid Save a grid image.
12--grid_rows Grid rows. 0 means automatic layout.
13--grid_cols Grid columns. 0 means automatic layout.
14--seed Random seed. If omitted, a random seed is generated.
15--device auto, cuda, or cpu.
16--allow_tf32 Enable or disable TF32 on compatible NVIDIA GPUs.
17--save_config Save generation_config.json.
18--no-save_config Do not save generation_config.json.
19--list_labels Print available label info and exit.
20--label_limit Number of labels to preview with --list_labels.