Skip to content

GitLab

Switch branch/tag
  1. 08 Aug, 2021 1 commit
  3. 05 Aug, 2021 2 commits
  5. 24 Jul, 2021 1 commit
  7. 23 Jul, 2021 1 commit
  9. 21 Jul, 2021 1 commit
  11. 20 Jul, 2021 1 commit
  13. 15 Jul, 2021 1 commit
  15. 13 Jul, 2021 1 commit
  17. 09 Jul, 2021 1 commit
    • Reed Wanderman-Milne's avatar
      With float16, always use LossScaleOptimizer. · 21286f77
      Reed Wanderman-Milne authored 
      Before, it was too easy to accidentally forget to set runtime.loss_scale, which had to always be done if mixed precision is used, otherwise the model would converge to worse accuracy. Now, all that needs to be done to use mixed precision is to set runtime.mixed_precision_dtype=float16.

PiperOrigin-RevId: 383767033
      21286f77
  19. 24 Jun, 2021 1 commit
  21. 23 Jun, 2021 1 commit
    • Reed Wanderman-Milne's avatar
      Improve error message when certain flags are not specified. · 0a9026e4
      Reed Wanderman-Milne authored 
      In nlp/train.py and vision/beta/train.py, certain flags are marked as required. Additionally, in certain functions, error messages are improved if a necessary flag is not specified, which is a fallback in case a file calling define_flags() does not mark the necessary flags are required. Previously if any of these flags were not specified, it would crash with a cryptic error message, making it hard to tell what went wrong.

In a subsequent change, I will mark flags as required in more files which call define_flags().

PiperOrigin-RevId: 381066985
      0a9026e4
  23. 22 Jun, 2021 1 commit
  25. 20 Jun, 2021 1 commit
  27. 16 Jun, 2021 1 commit
  29. 11 Jun, 2021 1 commit
  31. 01 Jun, 2021 1 commit
  33. 28 May, 2021 1 commit
  35. 17 May, 2021 1 commit
  37. 14 May, 2021 1 commit
  39. 13 May, 2021 1 commit
  41. 06 May, 2021 1 commit
  43. 16 Apr, 2021 1 commit
  45. 13 Apr, 2021 2 commits
  47. 12 Apr, 2021 1 commit
    • Reed Wanderman-Milne's avatar
      Use nonexperimental mixed precision API for official models. · 0d8f9807
      Reed Wanderman-Milne authored 
      For all modified calls to set_mixed_precision_policy(), the loss_scale argument was removed, as it cannot be passed if the nonexperimental API is used. For all such callers, the loss_scale is later used to explicitly create a LossScaleOptimizer, so removing the argument has no impact.

Switching to the non-experimental LossScaleOptimizer has no effect, as it has near identical behavior and all isinstance checks within the official models check for the non-experimental version.

PiperOrigin-RevId: 368101975
      0d8f9807
  49. 05 Apr, 2021 3 commits
  51. 02 Apr, 2021 1 commit
  53. 01 Apr, 2021 1 commit
  55. 24 Mar, 2021 1 commit
  57. 19 Mar, 2021 1 commit
  59. 18 Mar, 2021 1 commit
  61. 17 Mar, 2021 1 commit
  63. 13 Mar, 2021 1 commit
  65. 03 Mar, 2021 2 commits
  67. 02 Mar, 2021 1 commit
  69. 01 Mar, 2021 1 commit