search

synesthesiam / voice2json

Command-line tools for speech and intent recognition on Linux
MIT License
1.09k stars 63 forks source link

no intent and slot value with the demo #79

Closed LeoMax-Xiong closed 1 year ago

LeoMax-Xiong commented 1 year ago

I use the demo and use the recognize-intent command to get the intent and slot, But i don't get the intent and slot. The final result as follows:

2022-10-15:18:20:01,632 DEBUG    [__main__.py:680] Loading profile from /root/.local/share/voice2json/en-us_kaldi-rhasspy/profile.yml
2022-10-15:18:20:01,638 DEBUG    [recognize.py:85] Loading /root/.local/share/voice2json/en-us_kaldi-rhasspy/intent.pickle.gz
{"text": "turn on the light", "intent": {"name": "", "confidence": 0}, "entities": [], "raw_text": "turn on the light", "recognize_seconds": 0, "tokens": [], "raw_tokens": [], "wav_seconds": null, "transcribe_seconds": null, "speech_confidence": null, "wav_name": null, "slots": {}}

how can I debug the graph and check the bug place. Thank you.

LeoMax-Xiong commented 1 year ago

I find that there is no intent graph during the training