diffusion-models-class VS diffusers

When you use stable diffusion, you can adjust the classifier free guidance scale to control how much it follows the input prompt. From what I understand(check https://github.com/huggingface/diffusion-models-class/tree/main/unit3), what cfg does is that it generates an unconditional image and an image conditional on the text prompt, and then scale up the difference.

https://huggingface.co/learn/nlp-course/ https://huggingface.co/docs/transformers (go through the task guide) https://github.com/huggingface/diffusion-models-class http://d2l.ai/ https://www.youtube.com/watch?v=VMj-3S1tku0&list=PLAqhIrjkxbuWI23v9cThsA9GvCAUhRvKZ

I couldn't understand the original paper(havent done meth in a long time). This blog post and short course help me to understand.

The Diffusion Models Course is another good resource to learn more technical details.

Next, you need to become familiar with the diffusion model. I recommend this huggingface's course(https://github.com/huggingface/diffusion-models-class) because it is very high quality and you will learn while using diffusers. At first glance, it may not seem directly related to this game, but in my case, knowing what is happening in diffusers helped me in many ways: trial and error, inspiration for ideas, etc. I had no knowledge of pytorch (the deep learning library used for diffusers), so I also took this course (https://www.udacity.com/course/deep-learning-pytorch--ud188) which was in the prerequisites for that huggingface's course. It was also very good.

Hugging Face Dreambooth Hackaton details

Details: https://github.com/huggingface/diffusion-models-class/blob/main/hackathon/README.md

I'm missing something about the random factor, from the sample code from https://github.com/huggingface/diffusers/blob/main/README.md

Hey people, I noticed that combining these two methods in a single forward pass increases the controllability of the generation quite a bit. I was kind of puzzled that sometimes ControlNet yielded better results than T2IAdapter for some cases, and sometimes it was the other way around, so I decided to test both at the same time, and results were quite nice. Some visuals and more motivation here: https://github.com/huggingface/diffusers/issues/5847 And it was already merged here: https://github.com/huggingface/diffusers/pull/5869

Open Parti Prompts: The better way to evaluate diffusion models (repo)

You have disabled the safety checker for by passing `safety_checker=None`. Ensure that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered results in services or applications open to the public. Both the diffusers team and Hugging Face strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling it only for use-cases that involve analyzing network behavior or auditing its results. For more information, please have a look at https://github.com/huggingface/diffusers/pull/254 .

For developers, TensorRT support also exists for the diffusers library via community pipelines. [1] It's limited, but if you're only supporting a subset of features, it can help.

In general, these insane speed boosts comes at the cost of bleeding edge features.

[1] https://github.com/huggingface/diffusers/blob/28e8d1f6ec82a6...

I was training sdxl UNET base model, with the diffusers library, which was going great until around step 210k when the weights suddenly turned back to their original values and stayed that way. I also tried with the ema version, which didn't change at all. I also looked at the tensor's weight values directly which confirmed my suspicions.

Merging LoRAs is essentially taking a weighted average of the LoRA adapter weights. It's more common in other UIs.

diffusers is working on a PR for it: https://github.com/huggingface/diffusers/pull/4473