Neural Network Architectures

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.

Architectures

• FFN
• RNN
• Seq2Seq
• GRU
• LSTM
• Transformer

For each of them, I have included some basic code for training and evaluation in the README.md file