Whisper Turbo in MLX
Project description
WTM (Whisper Turbo MLX)
This repository provides a fast implementation of the Whisper model using MLX, designed for efficient audio transcription.
Features
- Fast Audio Transcription: Optimized for quick processing.
- Simplified Architecture: Focus on ease of use with minimal decoding strategies.
- Custom Tokenizer: Supports multilingual tokenization.
Installation
# Quick install (note: PyPI version may not always be up to date)
brew install ffmpeg
pip install whisper-turbo-mlx
# For the latest version, you can install directly from the repository:
# git clone https://github.com/JosefAlbers/whisper-turbo-mlx.git
# cd whisper-turbo-mlx
# pip install -e .
Usage
To transcribe an audio file, call the transcribe function:
transcribe 'test.wav'
Contributing
Contributions are welcome! Feel free to submit issues or pull requests.
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
whisper_turbo_mlx-0.0.1rc7.tar.gz
(377.5 kB
view details)
Built Distribution
File details
Details for the file whisper_turbo_mlx-0.0.1rc7.tar.gz.
File metadata
- Download URL: whisper_turbo_mlx-0.0.1rc7.tar.gz
- Upload date:
- Size: 377.5 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | a43ce703af7ef043235b4846f9ada734d827946f411abde9a327fa6b2148e63e |
|
| MD5 | 1dc41c2ff56b1f010f2cc5852b7d8546 |
|
| BLAKE2b-256 | d1ab87581ed6a2a5d70bfc9eb651f0223999b7699a47d9a484b75c53957e68d3 |
File details
Details for the file whisper_turbo_mlx-0.0.1rc7-py3-none-any.whl.
File metadata
- Download URL: whisper_turbo_mlx-0.0.1rc7-py3-none-any.whl
- Upload date:
- Size: 377.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 4757555e3ecd4f465dcdb250aabf87a035f7bed2a4e7084e998736d60778189c |
|
| MD5 | c6313d62e50939e3f3e5e448a0f5b003 |
|
| BLAKE2b-256 | 38cc1fa63e778f44c2c506b5989fc6fe5d5c914815a9e1409c3930b1c097a9f4 |
