Skip to content

v1.4 馃悰 Bug fix

Choose a tag to compare

@WenjieDu WenjieDu released this 27 Apr 12:10
· 41 commits to main since this release

Fixed some reported bugs in this release.

馃憤 Kudos to our new contributors @haoyu-haoyu and @shaun0927!

What's Changed

  • Update docs by @WenjieDu in #824
  • Update docs by @WenjieDu in #825
  • Fix failed CI (cannot import name 'GenerationMixin' from 'transformers.generation') by @WenjieDu in #829
  • Fix failed greetings with 403 error on new PRs by @WenjieDu in #831
  • Update stale manager by @WenjieDu in #832
  • Correct arg typing error in ModerTCN by @WenjieDu in #834
  • fix: CUDA device compatibility for Mean/Median/Lerp imputers by @haoyu-haoyu in #820
  • fix: allocate CRPS accumulator on the input device by @shaun0927 in #837
  • build(deps): bump pypa/gh-action-pypi-publish from 1.13.0 to 1.14.0 by @dependabot[bot] in #835
  • fix: validate inputs and preserve numpy return type in calc_quantile_loss by @shaun0927 in #839
  • fix: keep TS2Vec classifier_type='linear_regression' working with DeprecationWarning by @shaun0927 in #840
  • fix: load checkpoints with weights_only when PyTorch supports it by @shaun0927 in #841
  • Fix some known issues by @WenjieDu in #842

New Contributors

Full Changelog: v1.3...v1.4