README.md

---
title: EcoSoundNet
emoji: 🔥
colorFrom: pink
colorTo: blue
sdk: streamlit
sdk_version: 1.41.1
app_file: app.py
pinned: false
---

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

# EcoSoundNet 🔥

Welcome to EcoSoundNet, a sound classification application built using Streamlit. This project leverages audio data from the **UrbanSound8K** dataset and implements machine learning models for sound classification. Check out the details below!

## Overview

EcoSoundNet uses state-of-the-art techniques for audio classification, aiming to detect various environmental sounds. The model is trained on the **UrbanSound8K** dataset, which contains 8,732 labeled sound excerpts from urban environments. The project includes preprocessing steps, model training, and validation with accuracy comparisons.

## Dataset

The dataset used for training the model is the [UrbanSound8K](https://www.kaggle.com/datasets/chrisfilo/urbansound8k) dataset. This dataset consists of sound clips from 10 different urban sound classes, making it suitable for environmental sound classification tasks.

## Notebooks for Preprocessing, Training, and Accuracy Comparisons

For detailed insights into the preprocessing, training, and validation steps, please refer to the following notebook:

[Training and Validation Notebook](https://colab.research.google.com/drive/1bOEBYO6emJOO20LYviUWK6krXbvoziIo?usp=sharing)

This notebook covers:
- Data preprocessing steps
- Model training
- Accuracy comparisons of different models

## Streamlit App

The main application is built using **Streamlit**, a powerful framework for creating data science applications. You can run the app using the file `app.py`.

### Configuration

- **SDK**: Streamlit
- **SDK Version**: 1.41.1
- **App File**: `app.py`
- **Pinned**: No