You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Diffusion pipeline: ene-to-end diffusion pipeline for inference. It already has a pretrained diffsion model, and then it can be used for inference. Pipeline stores all components (models, schedulers, and processors). Also provides model loading, downloading and saving.
Model: Pretrained diffusion model. It can be used for inference. Diffusers model built on the base class ModelMixin that is a torch.nn.module.
The model configuration is a 🧊 frozen 🧊 dictionary, which means those parameters can’t be changed after the model is created. Therefore, the model is always static and reproducible. We consider the other parameters.
Schedulers: In the diffusion process, the scheduler is operated. Scheduler also be called Samplers in other diffusion models implementations. The scheduler can control denoising steps, denoising speed, noise level, and other parameters. (quality trade-off)
How to run?
Set the environment
make env
conda activate 01-diffusers-quick-start
make setup
Pipeline example
python pipeline.py
Loading pipeline components...: 29%|██████████████████████████████████████████ | 2/7 [00:00<00:01, 3.08it/s]`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The value `text_config["id2label"]` will be overriden.
`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The value `text_config["bos_token_id"]` will be overriden.
`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The value `text_config["eos_token_id"]` will be overriden.
Loading pipeline components...: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:01<00:00, 4.57it/s]
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [00:09<00:00, 5.32it/s]
ls outputs
image_of_squirrel_painting.png