A complete webapp which performs neural style transfer on videos for multiple styles and enables the user to preview, download and save the generated videos!
- Frontend interface created using React for dynamic upload and preview of images and videos. Used Chicago Art API to enable text based image search. Added client side routing for a smooth experience with a youtube-like interface.
- Backend created using Flask for a powerful python backend which utilizes Tensorflow Fast Arbitrary Style Transfer model and pymongo for connection to mongodb to process requests.
- Used MongoDB for fast database query retrieval.
- Dockerized the entire application for scalability and portability.
- Clone the repository by running git clone
- Run
docker-compose up
- The application can now be accessed at localhost:5000