deblan-mirror/buzz
1
0
Fork 0
mirror of https://github.com/chidiwilliams/buzz.git synced 2026-03-14 22:55:46 +01:00
Code Issues Projects Releases Packages Wiki Activity

Adding note on CUDA 12 for Faster whisper (#875)

Browse source
This commit is contained in:
Raivis Dejus 2024-08-07 21:56:37 +03:00 • committed by GitHub
commit 62eb2096f6
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 18 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

7
buzz/transcriber/recording_transcriber.py
View file

@ -62,6 +62,9 @@ class RecordingTranscriber(QObject):
model_path = self.model_path
keep_samples = int(0.15 * self.sample_rate)
if torch.cuda.is_available():
logging.debug(f"CUDA version detected: {torch.version.cuda}")
if self.transcription_options.model.model_type == ModelType.WHISPER:
device = "cuda" if torch.cuda.is_available() else "cpu"
model = whisper.load_model(model_path, device=device)
@ -76,6 +79,10 @@ class RecordingTranscriber(QObject):
logging.debug("CUDA GPUs are currently no supported on Running on Windows, using CPU")
device = "cpu"
if torch.cuda.is_available() and torch.version.cuda < "12":
logging.debug("Unsupported CUDA version (<12), using CPU")
device = "cpu"
model = faster_whisper.WhisperModel(
model_size_or_path=model_path,
download_root=model_root_dir,

7
buzz/transcriber/whisper_file_transcriber.py
View file

@ -52,6 +52,9 @@ class WhisperFileTranscriber(FileTranscriber):
"Starting whisper file transcription, task = %s", self.transcription_task
)
if torch.cuda.is_available():
logging.debug(f"CUDA version detected: {torch.version.cuda}")
recv_pipe, send_pipe = multiprocessing.Pipe(duplex=False)
self.current_process = multiprocessing.Process(
@ -146,6 +149,10 @@ class WhisperFileTranscriber(FileTranscriber):
logging.debug("CUDA GPUs are currently no supported on Running on Windows, using CPU")
device = "cpu"
if torch.cuda.is_available() and torch.version.cuda < "12":
logging.debug("Unsupported CUDA version (<12), using CPU")
device = "cpu"
model = faster_whisper.WhisperModel(
model_size_or_path=model_size_or_path,
download_root=model_root_dir,

4
docs/docs/faq.md
View file

@ -35,10 +35,14 @@ sidebar_position: 5
On Windows see [this note](https://github.com/chidiwilliams/buzz/blob/main/CONTRIBUTING.md#gpu-support) on enabling CUDA GPU support.
For Faster whisper CUDA 12 is required, computers with older CUDA versions will use CPU.
6. **How to fix `Unanticipated host error[PaErrorCode-9999]`?**
Check if there are any system settings preventing apps from accessing the microphone.
On Windows, see if Buzz has permission to use the microphone in Settings -> Privacy -> Microphone.
See method 1 in this video https://www.youtube.com/watch?v=eRcCYgOuSYQ
For method 2 there is no need to uninstall the antivirus, but see if you can temporarily disable it or if there are settings that may prevent Buzz from accessing the microphone.