comfyanonymous
After runed “comfy update” on Ubuntu, then the image generated by z-Image is a noise map
Today , I just runed “comfy update” on Ubuntu, I will generate image by z-Image, but the image generated is a noise map. I can generated image by z-Image yesterday.
I can generate image by Pony model after comfyui updated now.
OS : Ubuntu GPU : RTX2080Ti-22G
After executing the command "git checkout v0.3.75" and force-reinstall, the system returned to normal.
I had the same issue. Do you use multiGPU by chance? I used VAEmultiloaderGPU for the vae and had the same problem. When i switched to the regular VaeLoader the problem ceased to exist. This all worked on 0.3.75.
I have the same card and its working. Try to add --fp16-unet to parameters.
i get the same output in comdyui (with z-image).
this is the setup: Python version: 3.11.8 (tags/v3.11.8:db85d51, Feb 6 2024, 22:03:32) [MSC v.1937 64 bit (AMD64)] ComfyUI version: 0.3.76 ComfyUI frontend version: 1.33.10
Total VRAM 11264 MB, total RAM 130956 MB pytorch version: 2.4.0+cu121 xformers version: 0.0.27.post2
Device: cuda:0 NVIDIA GeForce RTX 2080 Ti : native
in forge neo, z-image does work, using the same "nodes" (model, vae, text encoder) as in comfyui. the setup in neo is: pytorch version: 2.9.1+cu128
so, i'm guessing the (old) pytorch is a problem, at least for me.
i upgraded pytorch to pytorch version: 2.9.1+cu128 and xformers to xformers version: 0.0.33.post1. the problem is gone now, for me.