MLP

General description

Python implementation of 'Multi-Layers Perceptron' Algorithm using only Numpy and Matplotlib.

Requirements

You should install numpy and matplotlib for clustering and visualization.

pip3 install numpy matplotlib

Usage

Generate Logic Gate Data

x,y = MLP.genLogicData('XOR')

Initialize Weights

weight_1 = MLP.dense(2,10)
weight_2 = MLP.dense(10,1)

Layer = MLP.layers(x,y,weight_1,weight_2)

Training Options

Option = MLP.option(ep = 10000,lr = 1,'sigmoid','gd',visualization = True)

Training

MLP = MLP()
MLP.Train(Layer,Option)

Test MLP

1. Open the runDemo file. You can just click F5 in an IDE environment to see the sample data results.

Author

Jongwon Kim : https://deep-eye.tistory.com/