What's Changed * Test Julia 1.9 by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/329 * Automated update to backend: v0.18.0 by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/331
Backend changes
[Diff since v0.17.1](https://github.com/MilesCranmer/SymbolicRegression.jl/compare/v0.17.1...v0.18.0)
- Overload ^ if user passes explicitly (MilesCranmer/SymbolicRegression.jl201) (MilesCranmer) - Upgrade DynamicExpressions to 0.8; LossFunctions to 0.10 (MilesCranmer/SymbolicRegression.jl206) (github-actions[bot]) - Show expressions evaluated per second (MilesCranmer/SymbolicRegression.jl209) (MilesCranmer) - Cache complexity of expressions whenever possible (MilesCranmer/SymbolicRegression.jl210) (MilesCranmer)
What's Changed * Highlight contributors by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/301 * Automated update to backend: v0.17.1 by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/320
What's Changed * Add paper 'Electron Transfer Rules of Minerals under Pressure…' by GCaptainNemo in https://github.com/MilesCranmer/PySR/pull/288 * Fix colab notebook example by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/295 * Add paper: "Data-Driven Equation Discovery of a Cloud Cover Parameterization" by agrundner24 in https://github.com/MilesCranmer/PySR/pull/302 * Pass through `enable_autodiff` parameter by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/316
New Contributors * GCaptainNemo made their first contribution in https://github.com/MilesCranmer/PySR/pull/288 * agrundner24 made their first contribution in https://github.com/MilesCranmer/PySR/pull/302
What's Changed * Complex-valued expressions by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/281 * Various fixes in backend (see https://github.com/MilesCranmer/SymbolicRegression.jl/releases/tag/v0.16.0)
What's Changed * Update backend version with warm start fix by MilesCranmer in https://github.com/MilesCranmer/PySR/pull/271 * This means that you can change the dataset or loss function, and `warm_start=True` will still work, and the losses will be re-computed.