Metadata-Version: 2.1
Name: lassonet
Version: 0.0.1
Summary: Reference implementation of LassoNet
Home-page: https://github.com/ilemhadri/lassonet
Author: Louis Abraham, Ismael Lemhadri
Author-email: louis.abraham@yahoo.fr, lemhadri@stanford.edu
License: MIT
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Operating System :: OS Independent
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Requires-Dist: torch
Requires-Dist: scikit-learn

# Feature Selection in Neural Networks

### Tips
LassoNet sometimes require fine tuning. For optimal performance, consider
- checking that the dense model with ![](https://latex.codecogs.com/svg.latex?%5Clambda%20%3D%200) has acceptable accuracy before running over the entire regularization path
- making sure the stepsize over the ![](https://latex.codecogs.com/svg.latex?%5Clambda) path is not too large. By default, the stepsize runs over the logscale between two values ![](https://latex.codecogs.com/svg.latex?%5Clambda_%7Bmin%7D) and ![](https://latex.codecogs.com/svg.latex?%5Clambda_%7Bmin%7D).

