This repository contains a deep learning solution for classifying 125x125 matrices in three-channel images of quarks and gluons impinging on a calorimeter. The task involves building and evaluating two different deep learning models: a VGG model with 12 layers and a ResNet-152 model.
The dataset is available at: Quark-Gluon Dataset
- Type: 125x125 matrices in three-channel images.
- Classes: Two classes of particles - Quarks and Gluons.
- Architecture: VGG with 12 layers.
- Modification: Reduced weights in the fully connected (FC) layers to speed up training and optimized architecture for classification.
- Architecture: ResNet-152.
- Details: Utilizes a very deep residual network with 152 layers to achieve high classification performance.
- Training Set: 80% of the data.
- Validation Set: 20% of the data.
- Both models are trained on the training set.
- Evaluated on the validation set to ensure no overfitting.
- The performance of each model is assessed based on classification accuracy.
- Model weights are saved and can be found in the
weights/
directory.
The code for training and evaluating the models is provided in the Jupyter notebook Quark_Gluon_Classification.ipynb
. The notebook includes:
- Data preprocessing and augmentation.
- Model definitions (VGG-12 and ResNet-152).
- Training loops.
- Evaluation and result visualization.
- Python 3.x
- PyTorch
- torchvision
- numpy
- matplotlib
- PIL
Install the required libraries using:
pip install torch torchvision numpy matplotlib pillow