Music-Genre-Classification

📄 Problem Statement

Description: Develop a machine learning model to classify music clips into genres such as rock, jazz, or classical using audio features like tempo or spectral characteristics.

Suggested Dataset: GTZAN Genre Collection or Free Music Archive (FMA).

Tools: Python, Librosa, Scikit-learn.

---

🎶 Music Genre Detection App

This is a web app built with Flask that lets users upload an audio file and returns the predicted music genre using audio feature extraction and machine learning.

🚀 Features

• Upload any audio file through a web form
• Automatic feature extraction using librosa
• Genre prediction using a LightGBM-trained model
• Deployed in Google Colab using ngrok

---

⚙️ Working

• Train LightGBM Model using the feature_30_sec.csv file
• Audio file can be uploaded using either Google Colab's files module or Web interface created using Flask and ngrok
• Uses librosa to extract features from audio like mfcc, chroma-stft, zero-crossing-rate, etc.
• Passes features to a trained model
• Returns the predicted music genre

Note: The app runs only while the Colab notebook is active. Create a config.py file with NGROK_TOKEN = 'Your_ngrok_Authotoken' within the directory structure.

---

📂 Contents

• README.md
• feature_30_sec.csv
• mgc_code.ipynb
• genres (directory)

Note: The genres folder contains audio files of different genres which can be used to test implementation.