Music Generation with RNN (LSTM)

This GitHub repository features a music generation project using Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) architecture. Trained on an Irish folk song dataset with ABC notations, the model learns patterns to generate unique musical sequences.

Sample Outputs: https://drive.google.com/drive/folders/1Qeh8aRggUsmKpXI0Vc18lO4gmjOHlWy-?usp=drive_link

Usage

1. Open music_gen.ipynb in Google Colab.
2. Connect to a GPU runtime for efficient training.
3. Execute cells sequentially to train the model.

Dependencies

• Python 3.7
• TensorFlow
• Keras
• abcmidi
• timidity

Install dependencies with:

pip install -r requirements.txt

Feel free to experiment with the notebook for music generation and contribute to the project. 🎶

Acknowledgments

MIT IntrotoDeeplearning for Datsset and Guidence. Feel free to contribute, report issues, or reach out with any suggestions or improvements. Happy music generation!