search

CoEDL / elpis

🙊 software for creating speech recognition models.
https://elpis.readthedocs.io/en/latest/
Apache License 2.0
152 stars 33 forks source link

Kaldi CLI fails on transcription #348

Open ginalevow opened 4 months ago

ginalevow commented 4 months ago

The Kaldi CLI fails when running 'transcribe.py' Traceback (most recent call last): File "/elpis/elpis/examples/cli/kaldi/transcribe.py", line 33, in t.prepare_audio(faudio) File "/elpis/elpis/engines/kaldi/objects/transcription.py", line 235, in prepare_audio self._process_audio_file(audio) File "/elpis/elpis/engines/kaldi/objects/transcription.py", line 63, in _process_audio_file self.audio_filename = audio.filename AttributeError: '_io.BufferedReader' object has no attribute 'filename'

I think the CLI example is out of sync with the use of werkzeug's FileStorage elsewhere.