What is this?

This i s a simple deep learning library

Usage?

Look at xor.py

What i have get with doing this library?

1. Principles of library design
2. Grasping how primitive version of PyTorch works
3. Solid understanding of backpropogation
4. Revealing some OOP concenpts like abstract classes and methods and inheritance
5. How using type annotations and mypy is cool

What else can be done?

1. Costum Tensor classe with GPU levarage
2. Cross-Entropy loss function
3. L1 L2 Regularization and dropout
4. More layers like conv, lstm
5. More activation functions. sigmoid, relu, softmax
6. More optimization tecniques, like Adam. Adding momentum.