WebApp-Multilingual-Speech-Recognition

A cutting-edge web application leveraging AI to perform multilingual speech recognition and transcription.

Table of Contents

• About the Project
• Features
• Technologies Used
• Prerequisites
• Usage
• License
• Acknowledgments

About the Project

This web application processes uploaded audio files and transcribes spoken language into text, supporting multiple languages. It is designed to provide high accuracy using Whisper. The app is ideal for generating accessible content, translating recorded meetings, or transcribing multilingual audio files.

Features

Speech-to-text transcription. Supports multiple languages. Intuitive and responsive web interface. AI-powered for high transcription accuracy.

Technologies Used

Frontend: HTML, CSS, JavaScript

Backend: FastAPI, WebSockets

AI Model: Whisper (fine-tuned version)

Other Tools: Python, Hugging Face, Lora, Bitsandbytes

Prerequisites

Python 3.9+: Ensure that Python is installed. You can download it from python.org.

Visual Studio Code: A lightweight and powerful editor with extensions like Python and Prettier. Download it from code.visualstudio.com.

Google Colab Pro : Use this for fine-tuning and training the Whisper model, offering enhanced computing resources.

Usage

Download the files to your local machine.

Launch the application in Visual Studio Code.

Choose a file to transcribe and upload it through the web interface.

View the transcription results once the file is processed.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• OpenAI Whisper
• FastAPI
• Hugging Face