Js Fakes 4bars

🚀 GPT-2 for Music

This project uses language models like GPT - 2 for music generation by representing music pieces as texts, reducing the task to language generation.

🚀 Quick Start

Language Models such as GPT - 2 can be utilized for Music Generation. The core concept is to transform music pieces into text, thus simplifying the task to Language Generation.

This model is a relatively small GPT - 2 instance trained on [TristanBehrens/js - fakes - 4bars](https://huggingface.co/datasets/TristanBehrens/js - fakes - 4bars). It generates 4 - bar Bach - like chorales with four voices (soprano, alto, tenor, bass) at a time.

If you're interested in contributing, want to say hello, or wish to learn more, you can find me on [LinkedIn](https://www.linkedin.com/in/dr - tristan - behrens - 734967a2/).

✨ Features

• Music Representation: Represents music pieces as texts for generation.
• Specific Output: Generates 4 - bar Bach - like chorales with four voices.

📚 Documentation

Model description

The model is GPT - 2 with 6 decoders and 8 attention - heads each. The context length is 512. The embedding dimensions are also 512. The vocabulary size is 119.

Intended uses & limitations

This model serves as a proof of concept, demonstrating that HuggingFace can be used for music composition.

How to use

There is a notebook in the repo that you can run on Google Colab.

Limitations and bias

Since this model has been trained on a very small corpus of music, it is overfitting heavily.

⚠️ Important Note

This model is just a proof of concept and overfits heavily due to the small training corpus.

💡 Usage Tip

You can run the notebook in the repo on Google Colab to use the model.