Implementing Backpropagation in My Deep Learning Framework

Built my own deep learning framework (e.g Tensorflow/PyTorch) to help me learn the basics better.

Implemented backprop and calculated gradients by myself. Results are comparable to using pytorch and tensorflow after training.

• Supports Relu and Sigmoid activation functions
• Supports BinaryCrossEntropy and Mean Squared Error Loss functions

Bonus

• implemented dynamic learning rate adjustment