TODO:
- Support addition of gradients during backward prop
- Support reduce functions like sum, max, min etc.
- Support matrix operations
- Train a NN on MNIST
- Figure out slicing
- Test every form of gradient using pytorch
- support saving and loading of models (using pickle)
- Support convolutions
- Train a CNN on CiFAR10
- Support Optimizer
- Make Modules
- CNN
- LSTM
- Transformer
Notes: