issues
search
accosmin-org
/
libnano
C++ numerical optimization and machine learning utilities using Eigen3
MIT License
4
stars
1
forks
source link
[bench] add benchmark for builtin models
#51
Open
accosmin
opened
2 years ago
accosmin
commented
2 years ago
configurable from command line (datasource, generator, loss, solver, model, tuner, splitter)
train the model and export results: CV, best parameters, evaluation metrics
should work for all linear and gradient boosting model variations
it should be possible to also configure what kind of weak learners to use