LoRA weights of Stable Diffusion v1-5 for fast image generation.
Important Usage Guidance
Use DDIM or Euler instead of LCM for sampling!
The name of each LoRA weights indicates how many inference steps they should be applied.
The name of each LoRA weights indicates whether they are able to use normal CFGs or small CFGs.
NormalCFG means that model equipped with the LoRA can use CFG value 4-12 for generation.
SmallCFG means that the model equipped with the LoRA can use CFG value 1-2 for generation.
Important Information
[[paper](https://arxiv.org/abs/2405.18407)]
[[code](https://github.com/G-U-N/Phased-Consistency-Model)]
[[huggingface](https://huggingface.co/wangfuyun)]
Description
FAQ
Comments (6)
I would prefer to be able to download them separately, not as one big zip file
@G_U_N Thank you, but I think maybe it is more helpful to have them each uploaded to Civitai as separate files, because tool such as Automatic 1111 can read the site and import helpful metadata that way.
@wes That's a good point. I will do that as your suggestion.
I have downloaded all the models that belong to 1.5 which is my favorite, I must tell you that you have done an excellent job, I did not work the small cfg, but the other excellent 16 steps has a quality that has little to envy the traditional system, even the 8 or 4 steps depending on what is needed, I do not understand how this post does not have more I like, keep it up
Amazing project!
I really liked to using it to speed up all the composition, especially with 8 or 16 steps.
P.s. I also found great this PCM (LCM_like_lora) and the 2-steps LoRa