- Compiled torch with CUDA support enabled for faster whisper processing for those who have NVidia GPUs
Windows version we have made available 2 versions, one compiled with python 3.8 and other with python 3.10. If you have issues with one please use the other one and drop and Issue report on the repo.
Linux version for distros with GLIBC 2.35 or newer. Compiled from Ubuntu 22.04
FOR macOS CATALINA AND ABOVE:
As the app is not downloaded from the AppStore, it may be necessary to allow it to run manually at
System Preferences > Security and Privacy
If you get a message "file damaged, you should move it to trash". You need to open a Terminal and type the following command to add an exception allowing the app to run:
xattr -cr Downloads/pyTranscriber*.app
Show your support
pyTranscriber is developed as a hobby and is available for free, but donations are welcomed and essential for further improvements.
Your donations will help me free up my time to do more direct work on pyTranscriber. If I can generate more income from sponsorship here, I can take time out of other jobs and commit more time to work on bug fixes, add new features, and make new releases, like possible future integration with Mozilla Deep Speech engine, which may improve accuracy and control over the speech detection.