Skip to content

GitLab

Switch branch/tag
  1. 09 Apr, 2021 1 commit
    • Reed Wanderman-Milne's avatar
      Remove dynamic_loss_scale argument to define_performance. · e353e4e5
      Reed Wanderman-Milne authored 
      All models which support loss scaling support dynamic loss scaling, so the argument has no purpose. It used to be that some models scaled the loss manually instead of using a LossScaleOptimizer, and so did not support dynamic loss scaling.

PiperOrigin-RevId: 367719521
      e353e4e5
  3. 08 Apr, 2021 1 commit
  5. 07 Apr, 2021 1 commit
  7. 06 Apr, 2021 2 commits
    • Hongkun Yu's avatar
      Clarify the deprecation warning for bert/ readme. · 8ccc242c
      Hongkun Yu authored 
      PiperOrigin-RevId: 367101911
      8ccc242c
    • Jeremiah Liu's avatar
      Disable temperature scaling during training. · fab47e9e
      Jeremiah Liu authored 
      For the `GaussianProcessClassificationHead`, the temperature scaling needs to be disabled during training to avoid unexpected modification to the learning rate, which harms model quality. (Unfortunately, this seems to require adding `training` to the `call` method).

Also set the default of `gp_cov_ridge_penalty` in `RandomFeatureGaussianProcess` to 1 to be consistent with that in the `GaussianProcessClassificationHead`.

PiperOrigin-RevId: 366917075
      fab47e9e
  9. 05 Apr, 2021 2 commits
  11. 01 Apr, 2021 5 commits
  13. 29 Mar, 2021 1 commit
  15. 26 Mar, 2021 3 commits
  17. 25 Mar, 2021 1 commit
  19. 24 Mar, 2021 1 commit
  21. 23 Mar, 2021 1 commit
  23. 22 Mar, 2021 4 commits
  25. 20 Mar, 2021 2 commits
  27. 18 Mar, 2021 1 commit
  29. 17 Mar, 2021 1 commit
  31. 16 Mar, 2021 3 commits
  33. 14 Mar, 2021 1 commit
  35. 12 Mar, 2021 3 commits
  37. 10 Mar, 2021 1 commit
  39. 09 Mar, 2021 1 commit
  41. 08 Mar, 2021 1 commit
  43. 05 Mar, 2021 1 commit
  45. 04 Mar, 2021 2 commits