with Apple MPS support for OpenAI Whisper
Project description
whisper-mps
An opinionated CLI to transcribe Audio files w/ Whisper on-device! Powered by MLX, Whisper & Apple M series
TL;DR - After our actual testing. The Whisper supported by MPS achieves speeds comparable to 4090!
80 mins audio file only need 80s on APPLE M1 MAX 32G! ONLY 80 SECONDS
🆕 Blazingly fast transcriptions via your terminal! ⚡️
We've added a CLI to enable fast transcriptions. Here's how you can use it:
Install whisper-mps
with pip
:
pip install whisper-mps
Run inference from any path on your computer:
# filename should be wav/mp3/mp4 etc
whisper-mps --file-name <filename>
Run inference from specfic model size:
# for example,with base model size, others:["tiny", "base", "small", "medium", "large"]
# Larger models require more loading time
# filename should be wav/mp3/mp4 etc
whisper-mps --file-name <filename> --model-name base
Run inference from YOUTUBE URL on your computer:
# default download behavior is to fetch the video as a mp3 file to save time
whisper-mps --youtube-url https://www.youtube.com/watch\?v\=jaM02mb6JFM
[!NOTE] The CLI is highly opinionated and only works on Apple MPS.
CLI Options
The whisper-mps
repo provides an all round support for running Whisper in various settings. More command-line support will be provided later
--file-name FILE_NAME
Path or URL to the audio file to be transcribed.
--model-name MODEL_NAME
size of the OPENAI Whisper model name, like tiny(default),base,small,etc
--youtube-url URL_ADDRESS
the youtube play url
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for whisper_mps-0.0.7-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 82ca1123a45ecad610f606c041b8159b30e8bbc1a5b17d86b8d0ef2da3f250af |
|
MD5 | 9e1bb3331e1676147c551b826d50c22b |
|
BLAKE2b-256 | 16ff73fc661cc9b011b0c8ea95ced6de61ed92336c5460edaf73592cee512d64 |