This is an implementation of the Neural Style Transfer algorithm using Pytorch.
The demo site can be viewed here.
In neural style transfer, a neural network (VGG16) takes one image, the "content image" (usually a photo), and "draws" it in the "style" of another image (usually a painting).
The neural networks used for styling images are implemented in Pytorch while the application itself is written using Streamlit.
For more details on the training code and training your own style models, refer to this repository.
To run the application locally, install the required dependencies by running
pip install -r requirements.txt
Once the requirements have been installed you can run the application using
streamlit run style-transfer/main.py
You can then go to http://localhost:8501 to view the running application.