DBGroup, Tsinghua University
-
@thu-ml, Tsinghua University
- Beijing, China
- https://jt-zhang.github.io/
Highlights
- Pro
Pinned Loading
-
thu-ml/SageAttention
thu-ml/SageAttention PublicQuantized Attention that achieves speedups of 2.1-3.1x and 2.7-5.1x compared to FlashAttention2 and xformers, respectively, without lossing end-to-end metrics across various models.
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.