summaryrefslogtreecommitdiff
path: root/candle-flash-attn/src/lib.rs
Commit message (Expand)AuthorAgeFilesLines
* chore: update flash attention kernels (#1518)OlivierDehaene2024-01-051-13/+421
* Fix for flash-attn. (#1310)Laurent Mazare2023-11-101-2/+2
* Properly set the is_bf16 flag. (#738)Laurent Mazare2023-09-041-6/+10
* BF16 support for flash-attn. (#737)Laurent Mazare2023-09-041-41/+81
* Add back the bf16 flash-attn kernels. (#730)Laurent Mazare2023-09-041-0/+2
* Relax the requirements on CustomOp. (#486)Laurent Mazare2023-08-171-2/+2
* Fix the flash-attention function names. (#282)Laurent Mazare2023-07-311-2/+2
* Flash attention without padding (varlen). (#281)Laurent Mazare2023-07-311-1/+230
* Add some flash attn test (#253)Laurent Mazare2023-07-261-10/+24
* Use bail rather than wrapping a string where possible. (#249)Laurent Mazare2023-07-261-2/+2
* Lining up the flash attn version with the non-flash one. (#248)Laurent Mazare2023-07-261-1/+18
* Again set a few extra params in flash-attn. (#245)Laurent Mazare2023-07-261-8/+8
* Proper flash-attn parameters. (#244)Laurent Mazare2023-07-261-7/+94
* Add flash attention (#241)Laurent Mazare2023-07-261-0/+59
generated by cgit v1.2.3 (git 2.39.1) at 2025-04-05 22:48:27 +0000