 juejuezi
|
e371bea04f
feat: change minimal supported CUDA version to 11.7 (#1206)
|
il y a 3 mois |
 Garrett Byrd
|
16025d8cc9
Clearer install instructions for CUDA and ROCm backends (#1147)
|
il y a 4 mois |
 Tri Dao
|
5ca83a9c71
Clean up softcapping bwd a bit
|
il y a 4 mois |
 rocking
|
d8f104e97a
Support AMD ROCm on FlashAttention 2 (#1010)
|
il y a 4 mois |
|
Tri Dao
|
7f67966cc7
FA3 initial code release
|
il y a 5 mois |
|
Tri Dao
|
da11d1b853
Bump v2.6.0
|
il y a 5 mois |
|
Tri Dao
|
320fb59487
Update citation
|
il y a 6 mois |
|
Tri Dao
|
2406f28805
Enable headdim 256 backward on consumer GPUs (Ampere, Ada)
|
il y a 9 mois |
 Tao He
|
204c3c6d1b
Fixes an error in comment (#785)
|
il y a 10 mois |
|
Tri Dao
|
54e80a3829
Implement page KV cache
|
il y a 10 mois |
 Erich Schubert
|
99ea4baa1d
Typo in README (#760)
|
il y a 11 mois |
|
Tri Dao
|
732654583c
Implement deterministic backward (thanks to Meituan)
|
il y a 11 mois |
|
Tri Dao
|
50d144c906
Mention Alibi in README
|
il y a 11 mois |
|
Tri Dao
|
7f31e7c16a
Bump to v2.3.2
|
il y a 1 an |
|
Tri Dao
|
5a83425442
Change constexpr int to constexpr static int
|
il y a 1 an |
|
Tri Dao
|
3a9fe7b0fa
Add change log
|
il y a 1 an |
|
Tri Dao
|
aa4fd2d166
Clarify that Windows is not supported right now
|
il y a 1 an |
|
Tri Dao
|
0c04943fa2
Require CUDA 11.6+, clean up setup.py
|
il y a 1 an |
 Jeffrey Quesnelle
|
1d817a8ffc
fix citation in README (#501)
|
il y a 1 an |
|
Tri Dao
|
45ba93cd96
Add newlines to README
|
il y a 1 an |
|
Tri Dao
|
9e5e8bc91e
Change causal mask to be aligned to bottom-right instead of top-left
|
il y a 1 an |
|
Tri Dao
|
d30f2e1cd5
Bump to v2.0.4
|
il y a 1 an |
 Ian Timmis
|
cbf982afa5
README syntax highlighting (#365)
|
il y a 1 an |
|
Tri Dao
|
d1a3b52f17
Add instruction about limiting number of ninja jobs
|
il y a 1 an |
|
Tri Dao
|
b4cc152e97
Make sure dout is contiguous
|
il y a 1 an |
|
Tri Dao
|
4f285b3547
FlashAttention-2 release
|
il y a 1 an |
|
Tri Dao
|
ce68305c84
Update installation instruction
|
il y a 1 an |
|
Tri Dao
|
f0c40b7ddb
Recommend Nvidia's Pytorch container
|
il y a 1 an |
|
Tri Dao
|
40a25c8ee7
Update roadmap
|
il y a 1 an |
 Anthony Hu
|
d63cfc3551
Use pyproject.toml to specify build dependencies
|
il y a 1 an |
