I think it is time that we can set attention for each inference without restart - Sage Attention breaking the Z Image generations
Open
Potential Bug
- Dominant language
- Python
- Stars
- 133k
- Forks
- 15.7k
- Avg merge
- 1d 7h
- Merged PRs (30d)
- 158
Description
With Sage Attention on RTX 5090 - around 18-20% faster
Without Sage Attention
Contributor guide
Assessment
This issue has not been assessed yet.