DaehanKimpushed 1 commit to dev • f8a43d3…4b0742e • on Dec 12, 2023
TPU support via shell script
DaehanKimpushed 1 commit to dev • 439264f…f8a43d3 • on Dec 10, 2023
DaehanKimforce pushed to dev • 29de7a3…439264f • on Dec 10, 2023
DaehanKimforce pushed to dev • 267f5e3…29de7a3 • on Dec 10, 2023
DaehanKimpushed 1 commit to dev • 62cde6c…267f5e3 • on Dec 10, 2023
DaehanKimforce pushed to dev • 1e55c53…62cde6c • on Dec 6, 2023
DaehanKimforce pushed to dev • 636d746…1e55c53 • on Dec 6, 2023
DaehanKimforce pushed to dev • 805b2ad…636d746 • on Dec 6, 2023
DaehanKimforce pushed to dev • 1296988…805b2ad • on Dec 6, 2023
DaehanKimforce pushed to dev • 68b342c…1296988 • on Dec 6, 2023
DaehanKimforce pushed to dev • c4cd4aa…68b342c • on Dec 6, 2023
DaehanKimforce pushed to dev • 6ab3696…c4cd4aa • on Dec 6, 2023
DaehanKimforce pushed to dev • f9a83f7…6ab3696 • on Dec 6, 2023
DaehanKimforce pushed to dev • 0cceb61…f9a83f7 • on Dec 6, 2023
DaehanKimpushed 1 commit to dev • 286dade…0cceb61 • on Dec 6, 2023
DaehanKimpushed 1 commit to main • 45eccb7…c775605 • on Dec 5, 2023
DaehanKimpushed 2 commits to main • 21d41e1…45eccb7 • on Dec 5, 2023
DaehanKimpushed 1 commit to main • e00658b…21d41e1 • on Apr 3, 2023
DaehanKimpushed 1 commit to main • fc5016d…3131f40 • on Apr 1, 2023
basic reward model training
Force push
basic reward model training
Force push
use bcewithlogitsloss instead of paper definition
DaehanKimpushed 2 commits to main • b4dc4df…6f3d797 • on Mar 15, 2023
add online and rejection-sampled datasets to alieviate overfitting
DaehanKimpushed 1 commit to main • b75da94…b4dc4df • on Mar 13, 2023
filter long sequences + add <|endoftext|> at ends of sequences
DaehanKimpushed 2 commits to main • 541c9ad…b75da94 • on Mar 13, 2023
reward model training available / ds_config setup for memory-efficien…
DaehanKimpushed 1 commit to main • 523ad52…541c9ad • on Mar 9, 2023
You can’t perform that action at this time.