ComfyUI-TiledDiffusion
ComfyUI-TiledDiffusion copied to clipboard
shiimizu
Tiled Diffusion, MultiDiffusion, Mixture of Diffusers, and optimized VAE
Hi and thanks you for a great library! I have an idea and don't know can I implement it or not :) I want to assign a new prompt for...
Hi, when trying to use the Tiled VAE Encode Node I get the following error: VAEEncodeTiled_TiledDiffusion Input type (float) and bias type (struct c10::BFloat16) should be the same Attached workflow....
  I scaled the image by 1.5x and replaced VAE Encode and VAE Decode with Tiled VAE Encode and Tiled VAE Decode. I didn’t change any other parameters, but...
so i just open up the example workflow provided, then switch the model to SDXL, with SDXL ControlNet, SDXL VAE, and change the latent size to match the recommended SDXL...
Hi, For whatever reason, when I use one or multiple controlnet with a mask + tiled diffusion (mixture of diffusion), the mask is applied on each tile:  Here the...
https://github.com/PRIS-CV/DemoFusion https://github.com/pkuliyi2015/multidiffusion-upscaler-for-automatic1111/blob/main/tile_methods/demofusion.py
Thanks for your work on tiled diffusion! Last month, a new upscaling technique was released called [DiffuseHigh](https://github.com/yhyun225/DiffuseHigh). I have been looking for an implementation in ComfyUI, but I haven't found...
It would be great to have support for Qwen-Image, so we can be free to generate large images. I have verified that both the tiled VAE encode/decode and diffusion nodes...
wan 2.2 use Tiled VAE Decode TypeError: VAEHook.__call__() got an unexpected keyword argument 'feat_cache'
Hi, First, thank you for creating such a game-changing tool. Its power is immense, but let's be honest: the vast ocean of custom nodes can be intimidating for many. We...
