Skip to content

Version 0.12.6
Compare
Choose a tag to compare
@mfeurer mfeurer released this 17 Apr 15:15
· 292 commits to master since this release
v0.12.6
275d0d6
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Verified
Learn about vigilant mode.

Version 0.12.6

  • ADD #886: Provide new function which allows fitting only a single configuration.
  • DOC #1070: Clarify example on how successive halving and Bayesian optimization play together.
  • DOC #1112: Fix type.
  • DOC #1122: Add Python 3 to the installation command for Ubuntu.
  • FIX #1114: Fix a bug which made printing dummy models fail.
  • FIX #1117: Fix a bug previously made memory_limit=None fail.
  • FIX #1121: Fix an edge case which could decrease performance in Auto-sklearn 2.0 when using cross-validation with iterative fitting.
  • FIX #1123: Fix a bug autosklearn.metrics.calculate_score for metrics/scores which need to be minimized where the function previously returned the loss and not the score.
  • FIX #1115/#1124: Fix a bug which would prevent Auto-sklearn from computing meta-features in the multiprocessing case.

Contributors v0.12.6

  • Francisco Rivera Valverde
  • stock90975
  • Lucas Nildaimon dos Santos Silva
  • Matthias Feurer
  • Rohit Agarwal
Assets 2
Loading