WorstNN

This is an attempt to make a Deep Learning Library from scratch for educational purposes.

Remember, this is just for understanding how NN works, I don't intend to compete with the likes of TensorFlow and PyTorch (as if I could, lol)

What to expect?

Implemented the following so far:

• Weight Initialization techniques such as Xavier, He and LeCun
• Different loss functions such as MSE, MAE, Huber Loss
• SoftMax Activation Layer
• Activation layers such as ReLU, Sigmoid, Tanh
• Xor Problem

More to come

I plan to add the following

• ReLU's variants such as Leaky ReLU, ELU, SELU
• Batch Normalization
• More loss functions such as Cross Entropy Loss
• Convolutional layer

How to use?

1. Clone the repository
   
2. Run the requirements (really just one requirement, numpy)
   
3. Run the following
   
4. python xor.py