1. 21 Jan, 2024 1 commit
  2. 20 Jan, 2024 1 commit
  3. 15 Jan, 2024 2 commits
  4. 14 Jan, 2024 5 commits
  5. 13 Jan, 2024 2 commits
  6. 12 Jan, 2024 1 commit
  7. 10 Jan, 2024 1 commit
  8. 08 Jan, 2024 1 commit
  9. 05 Jan, 2024 3 commits
  10. 04 Jan, 2024 1 commit
  11. 03 Jan, 2024 1 commit
  12. 26 Dec, 2023 1 commit
  13. 25 Dec, 2023 4 commits
  14. 24 Dec, 2023 3 commits
  15. 23 Dec, 2023 1 commit
  16. 22 Dec, 2023 7 commits
  17. 20 Dec, 2023 4 commits
    • Tri Dao's avatar
      Format flash_attn_interface.py · bc28eacc
      Tri Dao authored
      bc28eacc
    • Tri Dao's avatar
      [Gen] Remove minor dead code · 0a146185
      Tri Dao authored
      0a146185
    • Sanghun Cho's avatar
      Support alibi, by Sanghun Cho from Kakao Brain · e4f726fc
      Sanghun Cho authored
      
      
      * hard-code alibi in fwd
      
      * use params.h as hun_heads
      
      * hard-code alibi in bwd
      
      * add alibi on/off option
      
      * compute alibi_start, ratio outside of kernels
      
      * fix minor merge conflict
      
      * add test_alibi.py
      
      * change apply_alibi() location before masking
      
      * add alibi in splitkv kernel
      
      * fix backward func # of returns
      
      * add out-of-bound check in apply_alibi()
      
      * update test_alibi.py
      
      * update test_alibi.py for kvcache
      
      * simplify alibi parameter interface
      
      * fix performance issue
      by computing alibi outside of branch
      
      * update test_flash_attn_varlen_func() for left padding
      
      * implement alibi_slopes (b, nh) loading
      
      * optimize apply_alibi() a bit
      
      * update test cases for alibi_slopes loading
      
      * reflect stylistic comments
      
      * disable "seqlenq_ngroups_swapped" when using alibi
      
      ---------
      Co-authored-by: default avatarmonk.detective <monk.detective@kakaobrain.com>
      e4f726fc
    • Tri Dao's avatar
      cd089597
  18. 17 Dec, 2023 1 commit