VBVR version of wan2.2 quantized to int8 convrot,
Tested with a few different random LoRAs and a couple different lightning LoRAs and everything came out great and very, very fast.
Made this cause I didn't see another that wasn't simplified, etc.
command used:
python_embeded\Scripts\ctq.exe -i VBVR-wan2.2-comfy-high-bf16.safetensors -o Wan2.2-VBVR-HIGH-int8_convrot.safetensors --int8 --scaling_mode row --comfy_quant --save-quant-metadata --wan --convrot
High-Five! Enjoy
