Skip to content

Releases: Dao-AILab/flash-attention

v2.2.5

24 Sep 08:05
@github-actions github-actions
v2.2.5
0a1d03c

Choose a tag to compare

View all tags
v2.2.5 
Bump to v2.2.5
Loading

v2.2.4.post1

22 Sep 06:57
@github-actions github-actions
v2.2.4.post1
bff3147

Choose a tag to compare

View all tags
v2.2.4.post1 
Re-enable compilation for Hopper
Loading

v2.2.4

21 Sep 06:40
@github-actions github-actions
v2.2.4
229080b

Choose a tag to compare

View all tags
v2.2.4 
Bump to v2.2.4
Loading

v2.2.3.post2

18 Sep 05:17
@github-actions github-actions
v2.2.3.post2
799f56f

Choose a tag to compare

View all tags
v2.2.3.post2 
Don't compile for Pytorch 2.1 on CUDA 12.1 due to nvcc segfaults
Loading

v2.2.3.post1

17 Sep 23:40
@github-actions github-actions
v2.2.3.post1
c984208

Choose a tag to compare

View all tags
v2.2.3.post1 
Set block size to 64 x 64 for kvcache to avoid nvcc segfaults
Loading

v2.2.3

16 Sep 08:47
@github-actions github-actions
v2.2.3
8c8b4d3

Choose a tag to compare

View all tags
v2.2.3 
Bump to v2.2.3
Loading

v2.2.2

11 Sep 06:48
@github-actions github-actions
v2.2.2
08c295c

Choose a tag to compare

View all tags
v2.2.2 
Bump to v2.2.2
Loading

v2.2.1

06 Sep 09:20
@github-actions github-actions
v2.2.1
a1576ad

Choose a tag to compare

View all tags
v2.2.1 
Bump to v2.2.1
Loading

v2.2.0

05 Sep 18:34
@github-actions github-actions
v2.2.0
6d673cd

Choose a tag to compare

View all tags
v2.2.0 
Bump to v2.2.0
Loading

v2.1.2.post3

04 Sep 06:46
@github-actions github-actions
v2.1.2.post3
4976650

Choose a tag to compare

View all tags
v2.1.2.post3 
Set single threaded compilation for CUDA 12.2 so CI doesn't OOM
Loading