search

mljar / mljar-supervised

Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation
https://mljar.com
MIT License
3k stars 401 forks source link

Control all parameters of the learning #87

Closed pplonski closed 4 years ago

pplonski commented 4 years ago

Right now user can not control all parameters of the learning, for example, the number of minimum steps of the algorithm (see discussion in https://github.com/mljar/mljar-supervised/issues/86). Please add a method to set any parameter of the learning algorithm.

pplonski commented 4 years ago

Right now, models are learned till early stopping, no additional time constrains are applied because of performance decrease (up to x20 slower, with callback)