Using Buzz 1.4.4 on Windows 11 Prof with all current updates and NVIDIA GeForce GTX 1650 with 4GB RAM GDDR5 and Cuda installed
When running Speaker Identification, I consistently get an error and the task does not complete.
Transcription with Whisper.cpp and model Large-v3-turbo completed successfully earlier.
The transcribed audio file is about 70 minutes long.
Here is the log file for the Speaker identification. Please let me know if you also need the logs for the transcription or any other information.
[2026-04-03 17:59:52,693] speaker_identification_widget.on_identify_button_clicked:561 DEBUG -> Speaker identification: Starting identification thread
[2026-04-03 17:59:59,837] config.apply_error_handling_defaults:193 WARNING -> OneLogger: Setting error_handling_strategy to DISABLE_QUIETLY_AND_REPORT_METRIC_ERROR for rank (rank=0) with OneLogger disabled. To override: explicitly set error_handling_strategy parameter.
[2026-04-03 17:59:59,855] export_config_manager.generate_export_config:108 INFO -> Final configuration contains 0 exporter(s)
[2026-04-03 17:59:59,855] training_telemetry_provider.configure_provider:309 WARNING -> No exporters were provided. This means that no telemetry data will be collected.
[2026-04-03 17:59:59,855] default_recorder.__init__:120 DEBUG -> Initializing DefaultRecorder with no exporters, exporting is disabled
[2026-04-03 18:00:01,472] config.<module>:54 DEBUG -> PyTorch version 2.8.0+cu129 available.
[2026-04-03 18:00:04,055] speaker_identification_widget.run:183 DEBUG -> Speaker identification worker: Starting
[2026-04-03 18:00:10,290] speaker_identification_widget.run:216 DEBUG -> Speaker identification worker: Using device=cuda
[2026-04-03 18:00:10,290] connectionpool._new_conn:1049 DEBUG -> Starting new HTTPS connection (1): huggingface.co:443
[2026-04-03 18:00:10,889] connectionpool._make_request:544 DEBUG -> https://huggingface.co:443 "HEAD /MahmoudAshraf/mms-300m-1130-forced-aligner/resolve/main/config.json HTTP/1.1" 307 0
[2026-04-03 18:00:10,923] connectionpool._make_request:544 DEBUG -> https://huggingface.co:443 "HEAD /api/resolve-cache/models/MahmoudAshraf/mms-300m-1130-forced-aligner/2d856eb340893e274480dfb15a7b2a94d7ab7f84/config.json HTTP/1.1" 200 0
[2026-04-03 18:00:13,722] connectionpool._make_request:544 DEBUG -> https://huggingface.co:443 "HEAD /MahmoudAshraf/mms-300m-1130-forced-aligner/resolve/main/tokenizer_config.json HTTP/1.1" 307 0
[2026-04-03 18:00:13,738] connectionpool._make_request:544 DEBUG -> https://huggingface.co:443 "HEAD /api/resolve-cache/models/MahmoudAshraf/mms-300m-1130-forced-aligner/2d856eb340893e274480dfb15a7b2a94d7ab7f84/tokenizer_config.json HTTP/1.1" 200 0
[2026-04-03 18:00:13,822] connectionpool._make_request:544 DEBUG -> https://huggingface.co:443 "GET /api/models/MahmoudAshraf/mms-300m-1130-forced-aligner/tree/main/additional_chat_templates?recursive=False&expand=False HTTP/1.1" 404 64
[2026-04-03 18:00:13,854] speaker_identification_widget.run:262 DEBUG -> Speaker identification worker: Generating emissions
[2026-04-03 18:08:57,565] speaker_identification_widget.run:270 DEBUG -> Speaker identification worker: Emissions generated
[2026-04-03 18:09:43,185] speaker_identification_widget.run:385 ERROR -> Speaker identification worker: Error - e != <star>
Traceback (most recent call last):
File "buzz\widgets\transcription_viewer\speaker_identification_widget.py", line 294, in run
File "ctc_forced_aligner\ctc_forced_aligner\alignment_utils.py", line 65, in get_spans
assert seg.label == ltr, f"{seg.label} != {ltr}"
^^^^^^^^^^^^^^^^
AssertionError: e != <star>
[2026-04-03 18:09:43,213] speaker_identification_widget.run:393 DEBUG -> Speaker identification worker: Cleaning up resources
[2026-04-03 18:09:43,229] speaker_identification_widget.on_identification_error:604 ERROR -> Speaker identification error: e != <star>
[2026-04-03 18:09:43,229] speaker_identification_widget._on_thread_finished:595 DEBUG -> Speaker identification: Thread finished
[2026-04-03 18:09:43,229] speaker_identification_widget.on_identification_finished:628 DEBUG -> Speaker identification: Empty result received
[2026-04-03 18:10:02,304] translator.start:155 DEBUG -> Translation queue received stop signal
[2026-04-03 18:10:02,304] translator.start:185 DEBUG -> Translation queue stopped
First, thank you for the very useful software!
Using Buzz 1.4.4 on Windows 11 Prof with all current updates and NVIDIA GeForce GTX 1650 with 4GB RAM GDDR5 and Cuda installed
When running Speaker Identification, I consistently get an error and the task does not complete.
Transcription with Whisper.cpp and model Large-v3-turbo completed successfully earlier.
The transcribed audio file is about 70 minutes long.
Here is the log file for the Speaker identification. Please let me know if you also need the logs for the transcription or any other information.