Python Multilayer Perceptron (MLP)

A raw Multilayer Perceptron implementation in pure Python with forward propagation, backpropagation, and various activation functions.

Features

• Forward Propagation
• Backpropagation for supervised learning
• Multiple activation functions (Sigmoid, ReLU, Softmax)
• Optimization with SGD (with momentum)
• Loss functions (MSE)
• Training and evaluation utilities

Future Features

• More activation functions (Tanh, Leaky ReLU, etc.)
• Additional optimization algorithms (Adam, RMSProp, etc.)
• Regularization techniques (Dropout, L1/L2 regularization)
• Batch normalization
• Weight initialization strategies

Struct

src/
├── __init__.py      
├── activation.py    
├── layer.py         
├── network.py       
└── utils.py