This is no-where near ready for production and I'm going to keep working on it at some time.
The goal of this project was to generate music indefinitely from the initial starting point of a small clip of sound. Initialy I did a lot of reasearch considering the type of audio to use. I did see a lot of other projects out there using midi file type to generate music with success. But around this time Milad one of my teachers recommened that I look into mel-spectrograms which can be used to identify sounds and music genres. So I thought I'd try to use them tho generate audio as well. It doesn't work to well right now but I'll keep updating this and hopefully generate a working model that you'll be able to use soon.
Visit this link https://github.com/JohnminerIv/Tune_genorator/blob/master/model_notebook.ipynb to see how the model(s) were created.
Listen here to a sample here https://drive.google.com/file/d/1kR14fitCfO9qK7wYGS3wApXqT7gDHp6A/view?usp=sharing
Watch a demonstration here https://photos.app.goo.gl/MQhKm4JL1Mxfs9Dy7
