Django-based Audio transcription Tool

The purpose is to create a Audio transcription tool using Django framework

Installation

Prerequisites

Python 3.8+ - Download

Pytorch - Pipeline, preprocessing

pip - Python package manager

FFmpeg - Audio processing (required)

Git - Version control

Django 4.0+ - Web framework

HuggingFace API Key - For AI transcription models (Get API Key)

Clone and setup

Clone repository

git clone https://github.com/vivupadi/Audio_transcription.git

cd Audio_transcription

Create virtual environment

python -m venv venv

Activate virtual environment

• Linux/macOS: source venv/bin/activate

• Windows: venv\Scripts\activate

Install dependencies

pip install -r requirements.txt

Deployment

Architecture Overview

Component	Purpose
Django	Managing Frontend & Backend services
HuggingFace API	Model API for speech to text conversion
Audio Processing	Preprocessing audio
Render	Hosting the Django app on the cloud

Steps

The model used is Whisper 'tiny'. The audio file was resampled to 'Mono' channel & 16KHz. (Since most of the audio transcription models are trained on mono channel and 16KHz channels) CI-CD implemented using Github actions. The APP is hosted on Render free tier.

📄 License

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

⭐ Star this repo if you find it helpful! Made with ❤️ by Vivek Padayattil