| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2026-10-18 | 2.6 kB | |
| v2.7.1 source code.tar.gz | 2026-10-18 | 2.6 MB | |
| v2.7.1 source code.zip | 2026-10-18 | 2.7 MB | |
| Totals: 3 Items | 5.3 MB | 0 | |
2.7.1 (2026-10-18)
Bug Fixes
- accept weights of shape (n, 1) with targets of shape (n, 1) (1823414)
- early rejection of bad variable names (#1418) (e5db909)
- evaluate integer and boolean inputs to predict in floating point (1823414)
- export max and min with three or more arguments (1fd4114)
- export Piecewise-based operators to PyTorch and JAX (#1415) (1fd4114)
- install Julia extensions without Pkg auto-precompile (#1433) (dd2c6f6)
- leave caller DataFrames untouched in fit and predict (#1417) (1823414)
- let extra_torch_mappings override built-in mappings (1fd4114)
- use element-wise max and min in JAX export (1fd4114)
Performance Improvements
- allocation-free operator constraint check (SymbolicRegression.jl PR 762 (github.com)) (78566a7)
Dependencies
- require DynamicExpressions 2.13 (SymbolicRegression.jl PR 772 (github.com)) (78566a7)
- update backend to v2.7.1 (78566a7)
Documentation
- share the light/dark choice between pysr.ai and julia.pysr.ai (#1432) (cf31619)
