Single Image Super Resolution

A brief description of what the project does and who it's for.

Table of Contents

• Installation
• Usage
• Run Commands

Installation

Instructions for setting up the environment, including dependencies. We recommend creating a virtual enviroment.

# Clone the repository
git clone https://github.com/alonhelvits/Single-Image-Super-Resolution.git

# Navigate into the project directory
cd yourproject

# Install dependencies (if applicable)
pip install -r requirements.txt

Usage

Download the processed dataset folder from here : https://drive.google.com/drive/u/2/folders/125CbInmLhFyBo0fOgFij4nayQDvXt15Z

From the same link you can also download the CBSR model checkpoints to test the CBSR model pipeline.

• Project
  • All python files from this repo
  • dataset
    • combined_500
      • train
      • val
      • test
  • classifier_100_epochs.pth
  • man_made_800_model.pth
  • nature_800_model.pth

Run Commands

To run a full training of a baseline model:

python3 main.py --baseline --dataset_path dataset/combined_500

If you downloaded the 3 model checkpoints:

• classifier_100_epochs.pth
• man_made_800_model.pth
• nature_800_model.pth

You can run the test on the data:

python3 main.py --CBSR --dataset_path dataset/combined_500