search

MStypulkowski / diffused-heads

Official repository for Diffused Heads: Diffusion Models Beat GANs on Talking-Face Generation
Other
471 stars 33 forks source link

Too slow, it takes 4 minutes to reason about one second of audio #11

Closed c1a1o1 closed 1 year ago

c1a1o1 commented 1 year ago

Too slow, it takes 4 minutes to reason about one second of audio

ngbien83 commented 1 year ago

Yeah, slow, result after 4hrs render: https://youtube.com/shorts/srwIXNDzTo0?feature=share

MStypulkowski commented 1 year ago

Yes, the generation time is long. We didn't experiment with different samplers but feel free to try it.

@ngbien83 you used the model incorrectly. Please read the README section on how to use your own identity frames.

liutaocode commented 1 year ago

r 4hrs rend

Does it disappear gradually?