1. 17 May, 2022 2 commits
  2. 28 Apr, 2022 2 commits
  3. 11 Mar, 2022 1 commit
  4. 07 Mar, 2022 1 commit
  5. 19 Feb, 2022 1 commit
  6. 18 Feb, 2022 1 commit
  7. 16 Feb, 2022 1 commit
    • Sangkug Lym's avatar
      gradient accumulation fusion · 83b1e42f
      Sangkug Lym authored
      remove redundant linear layer class definition
      
      add fuse_gradient_accumulation attribute to weights for simple targetting
      
      reflect feedback and clean up the codes
      
      arg change
      83b1e42f
  8. 15 Feb, 2022 3 commits
  9. 11 Feb, 2022 1 commit
  10. 04 Feb, 2022 1 commit
  11. 01 Feb, 2022 3 commits
  12. 28 Jan, 2022 1 commit
  13. 27 Jan, 2022 1 commit
  14. 26 Jan, 2022 1 commit
  15. 25 Jan, 2022 1 commit
  16. 24 Jan, 2022 3 commits
  17. 22 Jan, 2022 1 commit
  18. 19 Jan, 2022 1 commit
  19. 13 Jan, 2022 1 commit
  20. 12 Jan, 2022 1 commit
  21. 04 Jan, 2022 1 commit
  22. 17 Dec, 2021 1 commit
  23. 24 Nov, 2021 1 commit
  24. 23 Nov, 2021 1 commit
  25. 22 Nov, 2021 1 commit
  26. 11 Nov, 2021 1 commit
    • Sangkug Lym's avatar
      persistent fused layer norm · a2fdcdf0
      Sangkug Lym authored
      fix the guard to fall back to the baseline fused layer norm kernel
      
      Persisten ln: move the guard for supported hidden sizes to layer norm module
      a2fdcdf0
  27. 10 Oct, 2021 1 commit
  28. 02 Sep, 2021 3 commits
  29. 23 Aug, 2021 1 commit
  30. 21 Aug, 2021 1 commit