# Hausa Speech Transcription
This repository contains a Streamlit application for transcribing Hausa audio files into text using a fine-tuned Whisper model. The model is trained specifically for the Hausa language, ensuring accurate transcription (Not a 100%) for this underrepresented language in NLP tasks.
---
## Features
- **Hausa Speech-to-Text**: Upload audio files in Hausa and get accurate transcriptions.
- **Supports Multiple Audio Formats**: WAV, MP3, OGG, and more.
- **Streamlit Web App**: An easy-to-use interface for seamless transcription.
- **Efficient Processing**: Built with optimized audio handling and resampling.
---
## Getting Started
### Prerequisites
- Python 3.8 or higher
- Install the required dependencies:
```bash
pip install streamlit torch transformers librosa numpy
git clone
github.com
cd hausa-whisper-transcription
streamlit run app.py