Skip to content

Releases: alan-turing-institute/autoemulate

v0.3.0

29 Apr 15:02
343864d
Compare
Choose a tag to compare

New features

  • History matching (#294)
  • Dimensionality reduction (#390)
  • Active learning (experimental) (#358)
  • End-to-end workflow example with a user-provided simulator (#423)
  • Initial PyTorch refactor (experimental)

What's Changed

New Contributors

Full Changelog: v0.2.1...v0.3.0

v0.2.1

30 Jan 16:27
Compare
Choose a tag to compare
  • fixed sci-kit learn dependency bug during installation
  • fixed warnings in tests

v0.2.0

27 Nov 17:08
Compare
Choose a tag to compare

see CHANGELOG for details.

v0.1.0.post1

02 Apr 16:15
Compare
Choose a tag to compare

Pre-release with basic functionality to automatically build emulators from simulation data, including:

  • 10 machine learning models commonly used in surrogate modeling, incl. neural nets
  • all models are scikit-learn compatible
  • cross-validation
  • hyperparameter optimisation (random and bayesian)
  • parallelisation
  • data pre-processing: scaling, dimensionality reduction
  • plotting