This project enables the local installation and use of WhisperX, an advanced audio transcription system based on OpenAI's Whisper but optimized for running on local hardware with or without a GPU. This project is made possible thanks to Whisperx and Faster Whisper. This document provides a general overview of the installation and links to the website where the complete installation and usage procedure for this project can be found.
- Miniconda
- CUDA (only for NVIDIA GPU users)
environment-cuda.yml
: Configuration file for automatic installation on systems with NVIDIA GPU.environment-cpu.yml
: Configuration file for automatic installation on systems without NVIDIA GPU.app.py
: Script to run the WhisperX user interface on Gradio.transcription_utils.py
: Transcription logic.
- Miniconda: Miniconda Installation
- CUDA: CUDA Installation (Only for NVIDIA GPU users)
- GitHub Repository: Download and setup of the repository. See details on our website.
For a detailed step-by-step installation process, please visit our website: View complete installation process
To use WhisperX after completing the installation:
- Activate the corresponding Conda environment.
- Run
python app.py
to start the Gradio user interface.