Skip to content

Latest commit

 

History

History
36 lines (26 loc) · 1.86 KB

CHANGE_LOG.md

File metadata and controls

36 lines (26 loc) · 1.86 KB

Change Log

SemiReward Update

  • [10/16/2024] SemiReward==0.2.0 has been released and SemiReward is accepted by ICLR'2024.
  • [10/18/2023] SemiReward==0.1.0 has been released, which is based on semilearn==0.3.1.
  • Update config files for SemiReward based on Pseudo-Labeling, FlexMatch, SoftMatch, and FreeMatch.
  • Update regression datasets, losses, evaluation hooks.

23/07/15/2023 USB Update

  • Added DeFixmatch algorithm.
  • Change the frequency of Tensorboard logging to every num_log_iter iterations.
  • Aim tracking now resumes tracking the same run when resuming training.
  • Simplified and improved the flexibility of the dependencies.
  • Fixes hard-coded repository path for Aim experiment tracking.

23/01/2023 USB Update

  • Added imbalanced Semi-Supervised setting and 8 imbalanced algorithms.
  • Added FreeMatch and SoftMatch algorithms.
  • Added registry mechanism for calling algorithms.
  • Added wandb and aimstack support.
  • Reformated 'train_step' structure.
  • Updated results (and logs) [Google Drive] [One Drive] with online wandb logs [Wandb].
  • Stable release semilearn==0.3.0
  • Updated Colab tutorial

16/10/2022 Update

  • Updated dataset download and pre-process instructions

18/09/2022 Update

  • Reformated 'train_step' structure with component hooks.
  • Added training hooks.