i did that because wan 2.2 on my pc is way more faster than qwen edit.
this is a tool, you need [get last frame] on your workflow to get the final result, then you can upscale this image with a standard upscale model.
WAN2.2UPSCALER_high_noise.safetensors