Commits · 6777336a1c7fabbd62637a9747d656f8ffdb2760 · gaoqiong / flash-attention

14 Jan, 2024 2 commits
- Move masking to a separate file (mask.h) · 6777336a
  Tri Dao authored Jan 14, 2024
  
  6777336a
- Remove seqq_parallel backward kernel that's not used · 9448264d
  Tri Dao authored Jan 14, 2024
  
  9448264d
13 Jan, 2024 1 commit
- Simplify writing softmax to gmem · a7b66ae2
  Tri Dao authored Jan 13, 2024
  
  a7b66ae2
12 Jan, 2024 1 commit
- Simplify SmemLayoutVtransposed in kernel_traits.h · 8d1b169e
  Tri Dao authored Jan 12, 2024
  
  8d1b169e
16 Sep, 2023 1 commit
- Implement rotary embedding in flash_attn_with_kvcache · ccbb14f3
  Tri Dao authored Sep 16, 2023
  
  ccbb14f3
04 Sep, 2023 1 commit
- Remove unused sdPsum in dot_do_o function · 5953c4f5
  Tri Dao authored Sep 03, 2023
  
  5953c4f5
01 Sep, 2023 1 commit
- Remove commented out code in bwd (#512) · 37e32feb
  Sophia Wisdom authored Sep 01, 2023
```
* Remove lots of comments

* Remove unused traits
```
  37e32feb
29 Aug, 2023 1 commit
- Implement splitKV attention · b1fbbd83
  Tri Dao authored Aug 29, 2023
  
  b1fbbd83
13 Aug, 2023 1 commit
- Prepare for Cutlass 3.2 · dbd79237
  Tri Dao authored Aug 13, 2023
  
  dbd79237
17 Jul, 2023 1 commit
- FlashAttention-2 release · 4f285b35
  Tri Dao authored Jul 17, 2023
  
  4f285b35