ShivamShrirao / diffusers

🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch
https://huggingface.co/docs/diffusers
Apache License 2.0
1.89k stars 505 forks source link

scheduler_config.json #202

Open rmac85 opened 1 year ago

rmac85 commented 1 year ago

This is more of a question than request...

I notice train_dreambooth.py has two different schedulers, a pipeline scheduler and noise scheduler, wondering if these two schedulers are reading from the same config file, since both appear to have different values, such as beta_start/end and DDPM also has variance_type. Can I add these settings manually to the line, like: noise_scheduler = DDPMScheduler.from_config(args.pretrained_model_name_or_path, subfolder="scheduler", beta_start=0.0001, beta_end=0.02, variance_type="learned")

Does this work or is it even necessary?