3.1.4

CustomRegressionModel and CustomClassificationModel added to algorithms
Convert sklearn like models to QSARtuna models
Convert preexisting QSARtuna models to uncertainty enabled versions
Automatic naming of models when no specific name supplied
Misc bug fixes
Relaxed Python versioning
Tracking/building latency improvements

3.1.3

• Improvements to model build process compute time and efficiency
• Model calibration reported at build time

3.1.2

This version incorporates AutoML into QSARtuna for automated model rebuilding. See the Documentation and README for more details.

3.1.1

-AmorProt and MAPC descriptors as co-variate auxiliary descriptors

• MAPC and Amorprot fingerprints added
• Bug fixes
• Precomputed descriptors inference improvements

3.1.0

This release has bug fixes and stability improvements.

PathFP fingerprints added
KNN algorithms added
ZScales descriptors and co-variate auxiliary descriptors
PCM support and peptide-protein modelling support
Precomputed descriptors inference improvements
ChemProp MPS support

3.0.0.1

QPTUNA was renamed to QSARtuna to avoid confusion with the original Optuna OSS
qsartuna-3.0.0.1.tar.gz

3.0.0

Uncertainty estimation, explainability, model calibration, ChemProp and more new functionailities

2.4.1

This is the first public release of Qptuna.