Replies: 1 comment 5 replies
-
Nice work! The main reference time is spent on model forward, so may I have a detailed time speed up here? |
Beta Was this translation helpful? Give feedback.
5 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Just a heads up:
If using diffusers, you can get a massive speedup by using Tiny AutoEncoder instead of the original VAE (40% faster on my 4060ti) for a minor trade-off in quality.
EDIT: This was a mistake on my part. See my comment below.
with VAE:
with TAE:
Beta Was this translation helpful? Give feedback.
All reactions