Incorrect generation of dictionary.txt

synesthesiam / rhasspy

Rhasspy voice assistant for offline home automation

https://rhasspy.readthedocs.io

MIT License

947 stars 101 forks source link

Closed mathquis closed 4 years ago

mathquis commented 4 years ago

During training (which fails silently), I get very strange entries in the dictionary.txt file in the profile/fr/kaldi/model/ folder.

adam a d A~
('adam', 2)(a d a m)

This looks like a format issue...

Also got in the logs:

ERROR (arpa2fst[5.5]:Input():kaldi-io.cc:756) Error opening input stream /profiles/fr/kaldi/model/data/lang/words.txt

The words.txt file is not there (re-downloaded the profile + kaldi). No changes.

synesthesiam commented 4 years ago

Got it! This must've been a mistake during refactoring. It'll happen when there's more than one pronunciation for a word. Fixed in master.