🎨 Neural Style Transfer with PyTorch

Apply the style of famous artworks to a content image using deep learning!

This project implements Neural Style Transfer (NST) using a pretrained VGG-19 network in PyTorch. You can run it through a Jupyter notebook or interactively with a responsive Streamlit app. 👉 Try the app live: neural-style-transfer-ryma.streamlit.app

---

🧠 Background

Neural Style Transfer is a computer vision technique that combines the content of one image with the style of another. It was introduced in:

A Neural Algorithm of Artistic Style
Leon A. Gatys, Alexander S. Ecker, Matthias Bethge (2015)
arXiv:1508.06576

The method uses a convolutional neural network (VGG-19) to extract features and optimize a new image that mimics the content and style characteristics.

This implementation is heavily inspired by the original code from:
👉 Yacine Mahdid's GitHub NST Tutorial

---

🗂️ Project Structure

├── app.py                   # Streamlit web app
├── notebook.ipynb          # Jupyter notebook
├── data/
│   ├── input/
│   │   ├── content/         # Content images
│   │   ├── styles/          # Style images
│   ├── output/              # Generated outputs organized by style
│   └── result/              # Visual summary (see above)

---

🚀 How to Use

🔬 Notebook (notebook.ipynb)

• Run the notebook on Kaggle (for free GPU support).
• Change the paths to use your own images if needed.
• It loops through multiple style images applied to one content image.

🌐 App (app.py)

Launch the interactive Streamlit app locally:

streamlit run app.py

or try it live 👉 neural-style-transfer-ryma.streamlit.app

• Upload your own content and style images.
• Adjust the number of optimization steps and style weight.
• Download the stylized result directly.

---

🖼️ Demo Screenshot

---

✅ Requirements

• torch, torchvision
• streamlit
• PIL, matplotlib

---

Feel free to explore, improve, and customize it for your artistic experiments!! 🎨✨