Faster Whisper Gradio

Real-time speech-to-text application using Faster Whisper with Gradio. This application utilizes DeeplX for translation. The application provides real-time transcription and translation, with different time windows for transcription.

Installation

1. Install PyTorch

Make sure to install the correct version of PyTorch with CUDA for GPU acceleration. Run the following command:

pip install torch==2.2.1+cu121

Install Faster Whisper

Install Faster Whisper directly from the GitHub repository:

pip install --force-reinstall "faster-whisper @ https://github.com/SYSTRAN/faster-whisper/archive/refs/heads/master.tar.gz"

This ensures you get the latest version of Faster Whisper for optimized real-time processing.

Running the Application

To start the Gradio interface for real-time speech-to-text transcription, execute the following command:

python app.py

This will launch the Gradio application.