ML_models

This repository contains my utilities and custom implementations for various machine learning models, including both original architectures and those based on research papers.

Current Uploaded Models

• u-net/: Implementation of the U-Net paper architecture for image segmentation.

• Simple_GAN/: Implementation of a GAN using only 1 layer nn (To show limitations and a brief idea of how GANs work).

• DCGAN: Implementation of a DCGAN with conv layers.

Notes

• Datasets are referenced in code but should be downloaded separately due to size.
• Dependencies for each project are listed in their respective requirements.txt files.

Getting Started

1. Clone the repository:

   git clone https://github.com/lexO-dat/ML_models.git

2. (Optional) Set up Git LFS for large model files:

   git lfs install
   git lfs pull

3. Install dependencies for the desired project:

   cd u-net
   pip install -r requirements.txt

4. Download datasets as needed (see code comments for dataset locations).