diff --git a/README.md b/README.md index e12efee..680db1e 100644 --- a/README.md +++ b/README.md @@ -2,7 +2,7 @@ [PhotoDoodle](https://github.com/showlab/PhotoDoodle) it a method about 'Learning Artistic Image Editing from Few-Shot Pairwise Data',you can use it in comfyUI # Update -* if use repo default use nf4 ,fast,and only 12G VRAM ,and not use mmgp,repo模式默认开启nf4量化,得益于新版的diffuser ,可以在不开启mmgp的情况下,轻松使用 +* add fp8 nf4 transformer and T5 quantinaze chocie,add seed,增加量化菜单,和T5量化选项,用于repo模式,增加seed。 # 1.Installation ----- @@ -33,7 +33,7 @@ pip install -r requirements.txt # 4 Example * if use single files,vae must choice "none" #flux1-dev-fp8.safetensors 16G 单体16G模型,内置clip和vae那种,vae必须选择"none",不开启mmgp,次要推荐使用 -![](https://github.com/smthemex/ComfyUI_PhotoDoodle/blob/main/assets/example.png) +![](https://github.com/smthemex/ComfyUI_PhotoDoodle/blob/main/assets/example0320.png) * if use repo 'black-forest-labs/FLUX.1-dev' or C:/youpath/black-forest-labs/FLUX.1-dev 如果使用repo可以用自动下载或本地,不开启mmgp,第一推荐使用 ![](https://github.com/smthemex/ComfyUI_PhotoDoodle/blob/main/assets/example_0317.png) * if use unet #11G 单体unet模型,没有内置clip和vae的,所以必须要连双clip和选vae,可以不开启mmgp,12G可能会OOM