Wenhao Chai

batch_size

1

It seems like the batch_size should only be 1 in both train and val, is that right?

merge with Dreambooth

1

What a wonderful work! I'm thinking about how this model merge with Dreambooth. I mean, how to train like Dreambooth style, just the same as vanilla stable diffusion is fine?

self.split_input_params NOT DEFINED

1

Hi, I found that in ./ldm/models/diffusion/ddpm.py L965: `if hasattr(self, "split_input_params"):`, but I can't find any definition of `split_input_params` in code/configs. Is it in TODO list or should be del?

### What's the feature? Hi! I would like to know do u have the plan to release the training code in diffusion model? ### Any other context? _No response_

New paper model release

2

Hi! I wonder if you have plan to share model weight of your new paper at huggingface?

DDP training support

2

Hi can u provide a brief insturction by using srun doing ddp training? Thank. you for your effort!

hi, will u guys do RLHF, DPO or self-reward training in xtuner? thanks for your great contribution to coMmunity!

Please forgive us if we did not involve your paper in this version as there are plenty of papers in recent years. Forgetting or accidents happen. You can send emails...

Please forgive us if we did not involve your paper in this version as there are plenty of papers in recent years. Forgetting or accidents happen. You can send emails...