Skip to content

upender8096/Generative-modelling-case-study

Repository files navigation

🍕 Generative Modelling Assignment – VS Code Project

This project is a VS Code-ready implementation of GANs across multiple domains, including synthetic data, medical imaging, cybersecurity, and creative AI.

📂 Project Structure

gan_vscode_project/ ├── .vscode/ ├── data/ │ ├── categories.txt │ ├── pizza.npy │ └── Wednesday-workingHours.pcap_ISCX.csv ├── notebooks/ ├── outputs/ │ ├── bloodmnist/ │ ├── cicids/ │ ├── part1/ │ └── quickdraw/ ├── src/ │ ├── init.py │ ├── common.py │ ├── part1_synthetic_gan.py │ ├── part2_bloodmnist_dcgan.py │ ├── part2_cicids_tabular_gan.py │ └── part2_quickdraw_dcgan.py ├── requirements.txt ├── run_all.py ├── setup.ps1 ├── setup.sh └── README.md


⚙️ Setup Instructions

Windows

python -m venv .venv ..venv\Scripts\Activate.ps1 pip install -r requirements.txt

macOS / Linux

python3 -m venv .venv source .venv/bin/activate pip install -r requirements.txt


Before Running the project add the datasets Wednesday-workingHours.pcap_ISCX.csv ====https://www.kaggle.com/datasets/chethuhn/network-intrusion-dataset pizza.npy===https://storage.googleapis.com/quickdraw_dataset/full/numpy_bitmap/pizza.npy (https://github.com/googlecreativelab/quickdraw-dataset)

🚀 Running the Project

Part 1

python src/part1_synthetic_gan.py

Part 2.1

python src/part2_bloodmnist_dcgan.py

Part 2.2

python src/part2_cicids_tabular_gan.py --input_csv data/Wednesday-workingHours.pcap_ISCX.csv

Part 2.3

python src/part2_quickdraw_dcgan.py --input_npy data/pizza.npy


▶️ Run All

python run_all.py


📊 Outputs

All outputs are saved in: outputs/

Includes:

  • Generated images
  • Loss graphs
  • PCA plots
  • Metrics

💡 Notes

  • Use fewer epochs for testing
  • Increase epochs for final results
  • Use generated outputs in report

⚠️ Troubleshooting

If torch not found: pip install torch torchvision torchaudio

If file path error: Check: data/pizza.npy data/Wednesday-workingHours.pcap_ISCX.csv


✅ Summary

This project demonstrates how GANs can learn patterns and generate synthetic data across multiple domains.

About

DCGAN model to generate synthetic pizza sketches from the QuickDraw dataset with visual and quantitative evaluation.

Resources

License

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors