1. 29 Jul, 2022 1 commit
  2. 28 Jul, 2022 1 commit
  3. 26 Jul, 2022 1 commit
  4. 25 Jul, 2022 1 commit
  5. 24 Jul, 2022 1 commit
  6. 21 Jul, 2022 1 commit
    • A. Unique TensorFlower's avatar
      Fix dropout rate bug. · b8234e65
      A. Unique TensorFlower authored
      Fix attention dropout rate bug. The output_dropout rate was used for attention dropout by mistake.
      
      PiperOrigin-RevId: 462470287
      b8234e65
  7. 14 Jul, 2022 1 commit
  8. 13 Jul, 2022 2 commits
  9. 30 Jun, 2022 1 commit
  10. 29 Jun, 2022 1 commit
  11. 27 Jun, 2022 1 commit
  12. 21 Jun, 2022 1 commit
  13. 20 Jun, 2022 1 commit
  14. 09 Jun, 2022 1 commit
  15. 08 Jun, 2022 1 commit
  16. 07 Jun, 2022 1 commit
  17. 21 May, 2022 1 commit
  18. 14 May, 2022 1 commit
  19. 04 May, 2022 2 commits
  20. 03 May, 2022 1 commit
  21. 20 Apr, 2022 1 commit
  22. 13 Apr, 2022 1 commit
  23. 28 Mar, 2022 1 commit
  24. 23 Mar, 2022 1 commit
  25. 11 Mar, 2022 3 commits
  26. 04 Mar, 2022 1 commit
  27. 03 Mar, 2022 1 commit
  28. 25 Jan, 2022 1 commit
  29. 19 Jan, 2022 1 commit
  30. 18 Jan, 2022 1 commit
  31. 04 Jan, 2022 1 commit
  32. 29 Dec, 2021 1 commit
    • Xin Wang's avatar
      Added block diagonal feedforward layer. · a7e60974
      Xin Wang authored
      This layer replaces the weight matrix of the output_dense layer with a block diagonal matrix to save layer parameters and FLOPs. A linear mixing layer can be added optionally to improve layer expressibility.
      
      PiperOrigin-RevId: 418828099
      a7e60974
  33. 21 Dec, 2021 1 commit
  34. 16 Dec, 2021 1 commit
  35. 03 Dec, 2021 1 commit
  36. 30 Nov, 2021 1 commit