Particle Classification with ResNet15

This repository contains code for training and evaluating a custom ResNet15 model for classifying particle types based on their hit energy and time matrices. The particles are classified into two categories: Electrons and Photons. The code includes data generation, preprocessing, model definition, training with cross-validation, and evaluation.

Table of Contents

1. Overview
2. Requirements
3. Data Generation
4. Data Preprocessing
5. Model Definition
6. Training and Validation
7. Testing
8. How to Use
9. Results
10. License

Overview

The code is designed to:

• Generate synthetic data representing hit energy and time matrices for Electrons and Photons.
• Preprocess and augment the data.
• Define and train a custom ResNet15 model with L1 regularization.
• Perform k-fold cross-validation to evaluate the model's performance.
• Evaluate the model on a separate test dataset and generate predictions for random test cases.

Requirements

To run this code, you need:

• Python 3.x
• PyTorch
• NumPy
• scikit-learn
• Matplotlib

You can install the required libraries using pip:

pip install torch numpy scikit-learn matplotlib