AssertionError: Input and weight inner dimensions must match
#64
by srgtsrgtrs - opened
Error with the official comfyui template, with the models all grabbed from this huggingface repo. shrugging my shoulders here.
Also the default comfyui workflow for LTX 2.5 is feeding a 1368x768 image into the resize image/mask node up to 1536 at longest length, where the actual final video size in the workflow is set to 0.9MP(?)
It's already a confusing workflow as is but i don't think the math even maths up lol