Adds multiple layer Neural Network

ugm2 / Deep-Learning-Workshop

Implementation of some of the main Deep Learning Algorithms with some examples

0 stars 0 forks source link

Adds multiple layer Neural Network #3

Closed ugm2 closed 3 years ago

ugm2 commented 3 years ago

This PR brings in a Neural Network with flexible number of layers, number of neurons and activation functions.

Things to be done left:

[X] Fix nn issue not learning properly
[x] Add load and save
[x] Include test metrics in examples
[x] Add cat vs none cat example for logistic regression too
[x] Generalize how examples are organized
[x] Create multi class classification example and include softmax activation function
[x] Include explanation of algorithms and examples in Readme

ugm2 commented 3 years ago

We should open issues for including tests. Also, I know we have exploding gradients in multiclass classification in the neural network. We should create anothre issue for fixing that problem, maybe using regularization