Skip to content

Pull requests: Dao-AILab/flash-attention

New pull request New
63 Open 236 Closed
63 Open 236 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Windows compile fix for MSVC
#1716 opened Jun 14, 2025 by loscrossos Loading…
Theoretically make compiling from pip quicker
#1703 opened Jun 8, 2025 by whrit Loading…
fix: fa3 backward check qkv with qkv_scale and dqkv
#1686 opened May 29, 2025 by yuyu5333 Loading…
1
[skip ci] libtorch agnostic FA3 north star proposal
#1685 opened May 28, 2025 by janeyx99 • Draft
Fix/deterministic dk dv
#1678 opened May 26, 2025 by yuWeiCute Loading…
8
Fix a bug in flash_attn_triton.py
#1668 opened May 15, 2025 by AminDarabi Loading…
Useuful command to install flash faster on behamoth clusters
#1660 opened May 10, 2025 by sleepingcat4 Loading…
Fix typos in multiple files
#1655 opened May 8, 2025 by co63oc Loading…
Patch RPATH of compiled Linux library to locate PyTorch and CUDA libraries in virtual env
#1634 opened Apr 30, 2025 by sisp Loading…
1
feat: support to tile K and V separately in FA3 backward
#1626 opened Apr 28, 2025 by beginlner Loading…
add checks for zero elements input of triton LayerNorm impl
#1621 opened Apr 27, 2025 by Luciennnnnnn Loading…
1
Add PT compileable support for flash_attn_with_kvcache
#1592 opened Apr 14, 2025 by jataylo Loading…
5
feat: fa3 custom ops for compatibility with PT Compile
#1590 opened Apr 13, 2025 by zhangheng408 Loading…
6
Add compile support for flash_attn_with_kvcache
#1554 opened Mar 25, 2025 by ani300 Loading…
1
Support cuda 12.8.1 and SBSA wheels
#1507 opened Feb 25, 2025 by johnnynunez Loading…
Remove ninja runtime dependency
#1484 opened Feb 9, 2025 by kevmo314 Loading…
1
4
[BugFix] Fix a wrong reference to seqlen_k variable in the fwd_splitkv kernel
#1455 opened Jan 21, 2025 by muoshuosha Loading…
1
Add missing tests/__init__.py
#1405 opened Dec 20, 2024 by BioGeek Loading…
1
Support dedicated compile[For Research]
#1384 opened Dec 12, 2024 by AllenDou • Draft
1
Fix deprecation warnings
#1382 opened Dec 12, 2024 by rongou Loading…
wrap func into torch ops to avoid torch.compile graphbreaks
#1333 opened Nov 13, 2024 by kumarkrishna Loading…
Promote wheels as alternative to pip install flash-attn
#1297 opened Oct 25, 2024 by simonw Loading…
4
fix: in newer versions of triton, tl.dot should take as input only q …
#1288 opened Oct 21, 2024 by EdouardYvinec Loading…
flashattnvarlen support tree attention
#1188 opened Aug 30, 2024 by efsotr Loading…
4
the test_flash_attn.py it's actually in parent directory
#1167 opened Aug 21, 2024 by ArtificialZeng Loading…
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.