1. 16 Apr, 2019 1 commit
    • Abhi Sharma's avatar
      Fix gradient overflow issue during attention mask · 9e666aaa
      Abhi Sharma authored
      This fix is in reference to issue #382. GPT2 can now be trained in mixed precision, which I've confirmed with testing. I also tested unconditional generation on multiple seeds before and after changing 1e10 to 1e4 and there was no difference. Please let me know if there is anything else I can do to make this pull request better. Thanks for all your work!
      9e666aaa
  2. 15 Apr, 2019 9 commits
  3. 12 Apr, 2019 2 commits
  4. 11 Apr, 2019 5 commits
  5. 03 Apr, 2019 2 commits
  6. 01 Apr, 2019 1 commit
  7. 27 Mar, 2019 1 commit
  8. 26 Mar, 2019 1 commit
  9. 24 Mar, 2019 6 commits
  10. 18 Mar, 2019 4 commits
  11. 14 Mar, 2019 1 commit
  12. 13 Mar, 2019 1 commit
  13. 12 Mar, 2019 4 commits
  14. 09 Mar, 2019 2 commits