InfoQGAN

InfoQGAN is the quantum version of InfoGAN. This repository contains code for training and testing the InfoQGAN model compare to the QGAN, InfoGAN, GAN.

Getting Started

You will need to install the appropriate packages via requirements.txt.

Before getting started, create a 2d_runs and iris_runs folder in the root directory. This folder is used to store TensorBoard data and training logs, and it is ignored by git.

The model is small enough to run without a GPU.

Directory Structure

INFOQGAN/
├── data/                           # Data files
│   ├── 2D/                         # 2D training data
│   ├── IRIS/                       # IRIS augmentation data
├── modules/                        # Models & utilities
│   ├── ndtest/                     # ndtest module from https://github.com/syrte/ndtest
│   ├── Discriminator.py            # Discriminator model
│   ├── MINE.py                     # Mutual info estimator (MINE)
│   ├── Generator.py                # Classical generator
│   ├── QGenerator.py               # Quantum generator
│   └── utils.py                    # Utility functions
├── .gitignore                      
├── 2D_train.py                     # Train QGAN/InfoQGAN on 2D data
├── 2D_train_classical.py           # Train GAN/InfoGAN on 2D data
├── 2d_custom_shape.ipynb           # Custom 2D shapes notebook
├── iris_train.py                   # Train QGAN/InfoQGAN on IRIS data
├── iris_train_classical.py         # Train GAN/InfoGAN on IRIS data
├── iris_augment_evaluation.ipynb   # Evaluate IRIS augmentation performance
├── README.md                       # Project overview
├── requirements.txt                # Required packages (Python 3.10)

Requirements

• Python 3.10 or higher (This code is written based on Python 3.12)
• The required packages are defined in the requirements.txt file.

To install the dependencies, run the following command:

pip install -r requirements.txt

How to Run

If you want to train for 2D dataset (please make 2d_runs folder before you run):

python 2d_train.py --model_type InfoQGAN

or

python 2d_train_classical.py --model_type InfoGAN

If you want to train with IRIS (please make iris_runs folder before you run):

python iris_train.py --model_type InfoQGAN

or

python iris_train_classical.py --model_type InfoGAN