New lightx
Will this be implemented in merge? https://huggingface.co/lightx2v/Wan2.2-Lightning/tree/main/Wan2.2-T2V-A14B-4steps-lora-250928
I read this earlier too and hope for this. I think I'm a MEGA convert after seeing how VACE can take multiple frames for their start frame. I thought it the input was first frame only. So now long length looks maybe possible after using the community pinned workflow. I was trying to make my own extended video segments on my own and single frame is impossible to deal with. However, one of VACE's problems is movement. Everyone is so stiff and trying to exaggerate movement descriptions doesn't even seem to help. Everything shifts to a cartoon look even though specifying photorealistic, etc. I'm hoping that creating LoRAs of the character will help fix this.
No need to change, new version of Lightning works only with 2 samplers and also have worse quality then old Lighxt2v merged to AiO...
Authors still working at this lora so maybe it will be better some day.
oooooo new lightning LORAs! There is a "low noise" LORA which I have been using. I'm going to see how the new one looks.