search

jianfch / stable-ts

Transcription, forced alignment, and audio indexing with OpenAI's Whisper
MIT License
1.59k stars 176 forks source link

When I run transcribe_minimal with -fw it gives me an error and doesn't complete #302

Closed xrishox closed 9 months ago

xrishox commented 10 months ago

Loaded Faster-Whisper large-v2 model
Traceback (most recent call last): File "/home/anonymous/anaconda3/envs/stable-ts/bin/stable-ts", line 8, in sys.exit(cli()) File "/home/anonymous/anaconda3/envs/stable-ts/lib/python3.10/site-packages/stable_whisper/whisper_word_level.py", line 1606, in cli transcribe_method = getattr(model, transcribe_method) AttributeError: 'WhisperModel' object has no attribute 'transcribe_minimal'

it works as expected without -fw . is this a bug? is this the expected outcome? is there anything I can do to make these work together?

jianfch commented 10 months ago

That is expected. There is only one transcription method supported for faster-whisper models and it is equivalent to transcribe_minimal() on the normal models (i.e. both are wrappers for transcribe_any()).