AnimeStyle Image Translation (LoRA Fine-Tuning)

Project Overview

This project fine-tunes a pre-trained deep learning model using LoRA (Low-Rank Adaptation) to convert real images into anime-style images. The model is trained with optimized hyperparameters to enhance the quality of anime-style transformations.

Project Structure

project-root/
│── anime-env/             # Virtual environment for dependencies
│── data/                  
│   ├── augmented_data/    # Augmented images
│   ├── original_data/     # Original dataset
│   ├── test_dataset/      # Test images
│   ├── dataset.md         # Dataset links stored on google drive
│── experiments/           
│   ├── model_checkpoints/ # Trained model checkpoints
│   ├── anime-style_training_log.csv  # Training logs
│   ├── experiments.md     # Links of checkpoints and training logs
│── fine-tuned-unet-lora/  # Fine-tuned model weights
│── notebooks/             # Jupyter notebooks for eda
│── Outputs/               # Generated anime-style images
│── scripts/               
│   ├── Model_run.py       # Main script to run the model
│── src/                   # Source code 
│── test/                  # Testing script
│── config.json            # Configuration file
│── Dockerfile             # Docker setup for reproducibility
│── requirements.txt       # Dependencies

Installation & Setup

1. Clone the Repository

git clone https://github.com/your-org/your-repo.git
cd your-repo

2. Create a Virtual Environment (Optional)

python -m venv anime-env
source anime-env/bin/activate  # macOS/Linux
anime-env\Scripts\activate     # Windows

3. Install Dependencies

pip install -r requirements.txt

4. Train the Model

Run the training script with LoRA fine-tuning:

python scripts/Model_run.py --config config.json

5. Generate Anime-Style Images

Use the trained model for inference:

python scripts/Model_run.py --input path/to/image.jpg --output path/to/save.jpg

Training Details

• Model Used: Fine-tuned UNet-based model with LoRA.
• Dataset: Augmented dataset of anime-style and real images.
• Training Framework: TensorFlow / PyTorch (specify accordingly).
• Performance Metrics: SSIM, PSNR, FID scores for image quality evaluation.

Contribution Guidelines

1. Fork the repository.
2. Create a feature branch (git checkout -b feature-name).
3. Commit changes (git commit -m "Added feature").
4. Push to the branch (git push origin feature-name).
5. Submit a Pull Request.

Future Enhancements

• Optimize training for better anime-style generation.
• Implement real-time inference with optimized performance.
• Deploy as a web application.

---

Author: Amisha Singh