Skip to content

GitLab

Switch branch/tag
  1. 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
  3. 29 Jul, 2021 2 commits
  5. 27 Jul, 2021 2 commits
  7. 21 Jul, 2021 2 commits
  9. 14 Jul, 2021 2 commits
  11. 13 Jul, 2021 1 commit
  13. 12 Jul, 2021 8 commits
  15. 09 Jul, 2021 1 commit
  17. 11 Jun, 2021 5 commits
  19. 10 Jun, 2021 16 commits