VoiceWorker

VoiceWorker is a Streamlit-based application for transcribing audio files or recordings using the Faster Whisper model. This app supports multiple languages and allows users to choose from various pre-trained model sizes for transcription tasks.

Why Faster Whisper?

Faster Whisper is a highly optimized implementation of OpenAI's Whisper model. It offers:

1. Enhanced Speed: Faster Whisper processes audio data significantly faster than the original implementation, making it ideal for real-time or large-scale transcription tasks.
2. Resource Efficiency: It uses less memory and computational power, enabling transcription on devices with limited resources.
3. Customizability: The model provides more flexibility with compute types (e.g., float16), which further optimizes performance on compatible hardware.

Difference Between Faster Whisper and Whisper

Feature	Faster Whisper	Original Whisper
Speed	Faster processing and inference	Slower inference
Resource Usage	Lower memory and computational needs	Higher memory and resource demand
Flexibility	Customizable compute types (e.g., float16)	Fixed compute type (float32)
Scalability	Suitable for real-time transcription	Best for offline, batch processing

Faster Whisper is especially advantageous for interactive applications like VoiceWorker, where speed and efficiency are critical.

Features

• Transcribe audio or video files.
• Record live audio for real-time transcription.
• Support for over 90 languages.
• GPU acceleration with CUDA (if available).
• Flexible model size selection.

Installation

1. Clone this repository:

   git clone https://github.com/amgawishx/VoiceWorker.git
   cd VoiceWorker

2. Install dependencies:

   pip install -r requirements.txt

3. Run the application:

   streamlit run app.py

Usage

1. Launch the app with the command above.
2. Choose your desired Whisper model size and transcription language in the app interface.
3. Upload an audio/video file or use the built-in audio recorder.
4. View the transcription output directly in the app.

Requirements

• Python 3.8+
• Streamlit
• Faster Whisper
• PyTorch
• GPU with CUDA support (optional but recommended for faster processing)

Acknowledgments

• Faster Whisper for providing the optimized transcription model.
• Streamlit for the interactive user interface framework.

---

VoiceWorker simplifies audio transcription tasks with speed and efficiency. Feel free to contribute or raise issues on the GitHub repository.