v0.1.9
KoljaB/RealtimeSTTv0.1.9Jan 29, 2024by KoljaB
AI Summary
This release migrates to PyTorch's multiprocessing backend and introduces configuration parameters for device selection and computing precision. It also improves usability by making the transcription process interruptible and fixes a specific issue.
Key Highlights
- Switched to `torch.multiprocessing`
- Added `compute_type`, `input_device_index`, and `gpu_device_index` parameters
- Added `recorder.abort()` for interrupting `recorder.text()`
- Fixed issue #20
New Features
- compute_type parameter
- input_device_index parameter
- gpu_device_index parameter
- recorder.abort() method
Full Release Notes
- switched to torch.multiprocessing - added compute_type (#14), input_device_index (select input audio device) and gpu_device_index (select gpu device) parameters - recorder.text() interruptable with recorder.abort() - fix for #20