|
What does this mean? :( === Warteschlange wird gestartet === Starte Auftrag: unbenannt.wav Audioumwandlung... Sprecher:innen identifizieren... Transkription... Transcription worker exited unexpectedly (code 3221226505). === Warteschlangenverarbeitung abgeschlossen === |
Replies: 1 comment 6 replies
|
From what I googled so far exit code 3221226505 seems like a Windows specific memory error. Could you share some details about your system (OS, CPU, RAM) what noScribe version you are using. |
|
Prozessor Intel(R) Core(TM) i7-10750H CPU @ 2.60GHz (2.59 GHz) Edition Windows 11 Home NVIDIA GeForce GTX 1660 Ti with Max-Q Design Normally it works all just fine. Reinstalling usually did it, if there was a problem. |
|
Regarding the virus/trojan check this. But for 0.7 there shouldn't be a diarize.exe anymore. Your system should be good enough for noScribe, I feel. However, we have some similar issues with the new version with Linux. Could you start your noScribe from the Windows terminal like this: The windows version is bundled by @kaixxx and I am not totally sure, how to invoke this from Windows, but I think this should work. Maybe you have to cd into the directory where you installed noScribe first. |
I am pretty sure that this is the culprit. I've had trouble with the GTX 1XXX range of cards before. To confirm this, please switch to CPU-processing for the transcription stage:
If we have confirmed that the GTX 1660 is the issue, we can try if we get it working. |
|
Solved! Thank you. :) Hallo, noScribe ist bereit! Version 0.7, Kai Dröge 2025 Bitte wähle links eine Audiodatei und den Dateinamen, unter dem das fertige Transkript gespeichert werden soll. Hab Geduld, die Transkription braucht üblicherweise 2 bis 5 Mal so lang wie die Audiodatei ist. Bevor man die gesamte Datei transkribiert, sollte man die Einstellungen an einem kurzen Ausschnitt testen. (Start = 00.00.00 and Stop = 00.03.00 bspw. transkribiert nur die ersten drei Minuten.) Viel Erfolg! Gewählte Audiodatei(en): Transkript-Dateiname(n): === Warteschlange wird gestartet === Starte Auftrag: unbenannt.wav Audioumwandlung... Sprecher:innen identifizieren... Transkription... [Text] Transkription beendet. === Warteschlangenverarbeitung abgeschlossen === |
|
OK, that's fine. You can use noScribe like it is. If you have lots of interviews to transcribe and want to try your GTX 1660 again, make sure that your NVIDIA drivers are on version 570.65 or higher and update the CUDA toolkit to the latest version (a reboot is required afterwards). However, chances for success are not very high, although in theory, these card should still be supported. |
|
Tank you a lot! I will consider it. :) |
I am pretty sure that this is the culprit. I've had trouble with the GTX 1XXX range of cards before. To confirm this, please switch to CPU-processing for the transcription stage:
C:\Users\<username>\AppData\Local\noScribe\noScribe\config.ymlforce_whisper_cputo'True'(don't forget the quotation marks)If we have confirmed that the GTX 1660 is the issue, we can try if we get it working.
Trying the suggestion form @gernophil would also be helpful, so that we get a bit more information about the problem.