search

rushter / MLAlgorithms

Minimal and clean examples of machine learning algorithms implementations
MIT License
10.73k stars 1.78k forks source link

Add missing neural network optimizers #55

Closed rushter closed 5 years ago

rushter commented 5 years ago

Currently, we are missing:

SanthoshBala18 commented 5 years ago

I would like to work on this if it is okay.

rushter commented 5 years ago

Sure, you can work on this one.

rushter commented 5 years ago

@SanthoshBala18 Someone has made a PR with Adamax https://github.com/rushter/MLAlgorithms/pull/56 You can work on the NAdam.

PyExtreme commented 5 years ago

@rushter , I am new to open soource. Can I please take up anything?

dipanshu231099 commented 5 years ago

@rushter shall i create a PR for a classifier using python scikit i have made.... also i have made a regression for a single variable input i could create a pr for...please suggest (the regression code is entirely by me and does not involve any imports except some very basic)

rushter commented 5 years ago

@dipanshu231099 This repository contains implementations of machine learning algorithms from scratch. We don't use any machine learning algorithms from scikit-learn.

rushter commented 5 years ago

@rushter , I am new to open source. Can I please take up anything?

I don't have any ideas right now, but I will try to come up with something.

We already have two people who are working on the current issue, so for future readers: Please don't start working on this issue.

Skub123 commented 5 years ago

missing some algorithm.

PyExtreme commented 5 years ago

@rushter , there is an optimizer known as Adabound which is fast as adam and conveges like SGD. Recently, It could published in ICLR 2019. Could it be added to the implementation list? If yes, then I would love to work on this.

rushter commented 5 years ago

@PyExtreme Sure, please add it.

SudarshanHV commented 5 years ago

@rushter Would you like to add an implementation of Neural Networks from scratch, using MNIST datasets, with some proper comments to help users? I have recently implemented it from scratch, with the help of a book, so, I would like to contribute