hyperparameter optimization for Javascript

A Javascript Library for hyperparameter optimization

Helps you find the optimal hyperparameters (constraints, weights or learning rates) for your learning algorithms.

Written in Javascript

Now that Tensorflow, the most popular machine learning framework, has been released as a Javascript API , we can create machine learning models that run in the browser, and it's easy to see why using Javascript for machine learning is on the rise.

Can be used in 2 ways

Link hpjs in your html file from cdn, or install in your project with npm


Utilize multiple parameters and multiple search algorithms (grid search, random, bayesian)


Best hyperparameters for some sample models we made (click on the model names to see more):
Parameter Examples

Hpjs features multiple parameter expressions, including a random int expression

Tensorflow Integration

Find the best optimizer and number of epochs for a small tensorflow.js model

Getting started

There are two ways to get TensorFlow.js and hpjs: via script tags or installing from npmUsing script tagsThe below code can be directly copied and pasted into an html file


via NPMThe example below is in React/Webpack.





Copyright (c) 2024 Atanas & Martin Stoyanov