Nibha10/environmental-sound-classification-app

Environmental Sound Classification App

This Streamlit app classifies environmental sounds using a LightGBM model trained on the UrbanSound8K dataset.

Features

• —Dataset analysis and visualization
• —Model performance metrics
• —Live prediction capabilities
• —Interactive visualizations

Usage

1. 1.Navigate to the "Live Prediction" page
2. 2.Upload a WAV file
3. 3.Get real-time predictions with probability distributions

Model Performance

• —Accuracy: 88.32%
• —Model: LightGBM
• —Features: OpenSMILE eGeMAPSv02

Requirements

Check requirements.txt for full list of dependencies.