Tensor Network - Matrix Product States (MPS)

This repository implements an image classification pipeline based on Matrix Product States (MPS) applied to the MNIST dataset.

Project Structure

• src/: Core Python modules for the model, preprocessing, and utilities.
• notebooks/: Jupyter notebooks for training, testing, and visualization.
• data/: Local storage for dataset files (.npy).

How to use

1. Local environment: Ensure you have the required dependencies installed (NumPy, PyTorch/TensorFlow, etc.).
2. Execution: Open notebooks/model_devl.ipynb to view the training process and classification results. The notebook imports logic directly from the src/ directory.

Contacts

Developed for research purposes in collaboration with the University of Ljubljana.