Setup

1. Clone the Repository:
   • Use GitHub Desktop to clone this repo, or download it as a ZIP from GitHub.
2. Install Dependencies:
   • Install Python 3.x.
   • Run pip install -r requirements.txt in a terminal, or manually install the libraries listed in requirements.txt.
3. Download the Dataset:
   • The datasize size is too large to upload, so i have provided the public link of google drive containing the dataset of Gurugram and california seperately and in the /data section, i have uploaded some of the same images just for references.
   • The dataset is too large for GitHub. Download it from https://drive.google.com/drive/folders/1zWbM4kLnmiKWS5BNSb-9HKHmEbaY26e_?usp=sharing for california dataset and https://drive.google.com/drive/folders/1EPZAisUa1mTHzlw4hHGU9uhVi9W9MxXm?usp=sharing for gurugram dataset.

Usage

1. Train the Model:
   • Open GAN_Training_California/Gurugram and run it in your Python environment (e.g., VS Code, Jupyter Notebook, or Google Colab), please note that i have trained my model with the dataset stored in my google drive please update the links according to your environment.
2. Generate Layouts:
   • Outputs will be saved in the outputs/ folder after training. (for some sample outputs)
   • Again, you can find all the ouptuts from public links: https://drive.google.com/drive/folders/1lxKUf5lZvBNFYvCfqTguD-2AyORkZb4k?usp=sharing and https://drive.google.com/drive/folders/18VLYigv4jZRC2USXeEWK84Ii98Cp-DfJ?usp=sharing

Results

• Current outputs are noisy (see outputs/), lacking clear urban features.
• California dataset shows slight diversity compared to Gurugram due to structured layouts.

Paper

• Find the published paper in paper/IJMSRT25MAR040.pdf.

Authors

• Vandna, Rahul Raj Parida, Mayank Raj (School of Engineering and Technology, KR Mangalam University, Gurugram, Haryana)

License

This project is submitted under the supervision of KR MANGALAM UNIVERSITY.