Kijai

/

WanVideo_comfy_GGUF

like 115

Hi, thanks in advance for your help!

Do VACE modules need to use the same quantization level as the main model?

For example, if I’m using this main model:
Wan2.2-T2V-A14B-HighNoise-Q3_K_S.gguf

can I pair it with this VACE module:
Wan2_1-VACE_module_14B_Q4_K_M.gguf

or does the quantization have to match exactly?