tomesd VS diffusers

They show up to 5.4 times greater: you can see his results in the image on the github repo here: https://github.com/dbolya/tomesd

Another optimization extension you can use as well is token merging which has reported around 5.4x faster image generation.

It merges redundant tokens: https://github.com/dbolya/tomesd So it can make the generation slightly faster.

Mods patched - Tomesd - Pillow-SIMD - OpenCV-CUDA (WIP) - Removed some unused imports and startup checking - Improved performance with reduced VRAM usage (tested on txt2img only) - Added a new option to use external RealESRGAN with --external-realesrgan

Were the 40 it/s perhaps achieved with ToMe?

Done! Proposal

This feature comes from this project: https://github.com/dbolya/tomesd

GitHub - dbolya/tomesd: Speed up Stable Diffusion with this one simple trick!

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