Description
wan 2.2
FAQ
Comments (9)
KSamplerAdvanced error
Given groups=1, weight of size [5120, 36, 1, 2, 2], expected input[1, 32, 13, 72, 128] to have 36 channels, but got 32 channels instead
Disable sageattention nodes and the middle nodes which are the kijai ones and you will get it working
@gamingworld113948 Thanks for the suggestion, I'll check if I haven't already done that (which is very possible).
Disabled sageattention, pytorch is 2.8, same error even if I disable pytorch nodes, models are those from original workflow. I'm out of solutions. :(
Thank you for your support!
@ionelbolintineanu707 hey one more thing if you are using gguf the all nodes like load clip load model and load text encoder need to be the gguf ones please check that also sometimes u might be accidently using scaled or other versions do check that
@ionelbolintineanu707 try putting all loaders as gguf not the normal loaders
@ionelbolintineanu707 I was getting this error message when using a certain workflow. It was due to Comfyui not being updated. Had to end up doing a fresh install of Portable and it fixed the problem
Thank you all for your support and I apologize for the lack of response, it's justified. On the other hand, I'm waiting to replace my current video card with a more powerful one, and until I do, I won't pay too much attention to ComfyUI. I wish you all the best!
@Grim0r Thank you for your suggestion, already done that and crashing other things with no good result for Wan 2.2. I'm waiting for more inspiration and a new graphic card. :-)