1. 02 Mar, 2022 1 commit
  2. 01 Mar, 2022 1 commit
  3. 23 Feb, 2022 1 commit
  4. 22 Feb, 2022 1 commit
  5. 15 Feb, 2022 1 commit
  6. 10 Feb, 2022 2 commits
  7. 28 Jan, 2022 2 commits
  8. 27 Jan, 2022 1 commit
  9. 26 Jan, 2022 1 commit
  10. 22 Jan, 2022 1 commit
  11. 19 Jan, 2022 1 commit
  12. 12 Jan, 2022 1 commit
  13. 21 Dec, 2021 1 commit
  14. 17 Dec, 2021 1 commit
  15. 02 Dec, 2021 1 commit
  16. 26 Nov, 2021 1 commit
  17. 24 Nov, 2021 1 commit
  18. 01 Oct, 2021 1 commit
  19. 20 Sep, 2021 1 commit
  20. 19 Aug, 2021 1 commit
  21. 18 Aug, 2021 2 commits
  22. 17 Aug, 2021 2 commits
  23. 16 Aug, 2021 2 commits
  24. 30 Jul, 2021 1 commit
    • Deepak Narayanan's avatar
      Support for pipeline parallelism in T5 model · 46c74b4c
      Deepak Narayanan authored
      - Accumulate encoder hidden state gradient to handle skip connection
      - Correctly compute the number of layers in encoder / decoder for T5 model
      - Ensure e weights are initialized the same way in embeddings
      - Synchronize embedding gradients across encoder and decoder for T5 model
      - Support for checkpoint loading and saving
      46c74b4c
  25. 21 Jul, 2021 1 commit
  26. 12 Jul, 2021 1 commit
  27. 05 Jul, 2021 2 commits
  28. 02 Jul, 2021 1 commit
  29. 29 Jun, 2021 1 commit
  30. 03 Apr, 2021 1 commit
  31. 02 Apr, 2021 1 commit
  32. 24 Mar, 2021 1 commit
  33. 19 Mar, 2021 1 commit
  34. 08 Mar, 2021 1 commit