Is there any way to run T2V with 4 step inference?

#13
by neuralityAI - opened

It seems like the only option is to use the wan 2.1 fast lora but the output is really bad. Is there a way to get high quality T2V with fast inference or not?

The lightning Lora's don't work, they break the output if i load them with pipe.enable_lora.

Sign up or log in to comment