Intrusion Detection System using a custom Neural Network

This project implements a custom neural network using PyTorch and trains it on NSL-KDD dataset.

Setup and Installation

• Clone the repository.
• This project using uv for dependency and package management. Your machine must have it installed as a prerequisite. Once the repository is cloned, you can setup the project at once by running the command uv sync inside the directory.
• To run the training pipeline, run the command uv run src/train.py from the project root directory.
• You can set the hyperparameters and models to be tested in src/constants.py.
• To test the model, run the command uv run src/test.py from the project root directory.

The model trained with current hyperparameters is able to achieve a F1 Score of 0.7 on the test split.

Any kind of PRs and suggestions for improvements are extremely welcomed.