This repository contains implementations of various neural network architectures. These were developed from the ground up as part of my deep learning studies
For the majority of the architectures, I followed a process of understanding the underlying theory, deriving the necessary backpropagation equations, and then implementing them in code.
In the case of more complex architectures, I utilized either jax (leveraging its automatic differentiation capabilities) or PyTorch for implementation.
- FFN
- RNN
- Seq2Seq
- GRU
- LSTM
- Transformer
- CNN
For each of them, I have included some basic code for training and evaluation in the README.md file