Skip to content

README.md

Latest commit

 

History

History
59 lines (45 loc) · 1.57 KB

README.md

File metadata and controls

59 lines (45 loc) · 1.57 KB

Music Generator

A Python-based AI application that generates music using an LSTM-based RNN model trained on music data. This project provides an end-to-end solution to compose music and play it as audio, leveraging the power of deep learning and Gradio for an interactive user experience.

Features

  • Music Generation: Uses a trained RNN model to generate music in ABC notation.
  • Interactive Interface: Gradio-based interface for easy interaction.
  • LSTM/GRU Support: Configurable choice of LSTM or GRU for music generation.
  • Audio Conversion: Converts ABC notation into playable audio.

Installation

Prerequisites

  • Python 3.8 or later
  • Required Python packages (install via requirements.txt):
    • torch
    • gradio
    • numpy
    • abc2midi

Steps

  1. Clone this repository: 
    git clone https://github.com/nullHawk/music-generation
cd music-generation
  2. Install dependencies: 
    pip install -r requirements.txt
  3. Place the trained model checkpoint at checkpoint/model.pth.

Usage

Running the Application

Launch the Gradio interface:

python app.py

Interacting with the App

  • Open the Gradio interface in your browser.
  • Click the "Generate Music" button to compose a new song.
  • Listen to the generated audio directly within the interface.

Example Output

  1. Click the Generate Music button.
  2. The app generates an ABC-notation music sequence using the RNN model.
  3. The ABC notation is converted to playable audio.
  4. Listen to the generated music in the interface.