This project is an experiment to try out different kinds of deep learning models for music. The main focus is to implement audio waveform models and symbolic music models, and use such models in the common context of creating interfaces for music making. The project consists of a backend which provides functionality to work with deep learning models for music, and an interface that lets us interact with these models using music hardware (e.g. MIDI controller) or a computer keyboard.
You can find the deployed application here - https://ai-music-maker.herokuapp.com/
This project is currently under development, so its pretty buggy. I will try working on it in my spare time.