Snowboy public model not working

Help
1

Hello,

i’ve encountered a strange issue. If I want to use jarvis.umdl, Rhasspy won’t work after restart. I can’t connect to the website.

# This site can’t be reached

**192.168.xxx.xxx**  refused to connect.

When I use my custom wakeword jarvis.pmdl it somehow works. Even renaming the jarvis.umdl to snowboy.umdl doesnt work.

profile.json (working):

{
“microphone”: {
“arecord”: {
“device”: “hw:CARD=Dummy,DEV=0”
},
“system”: “arecord”
},
“mqtt”: {
“enabled”: true,
“host”: “192.168.178.19”
},
“sounds”: {
“error”: “${RHASSPY_BASE_DIR}/sounds/wav/error.wav”,
“recorded”: “${RHASSPY_BASE_DIR}/sounds/end_of_input.wav”,
“wake”: “${RHASSPY_BASE_DIR}/sounds/start_of_input.wav”
},
“speech_to_text”: {
“system”: “kaldi”
},
“text_to_speech”: {
“system”: “picotts”
},
“wake”: {
“snowboy”: {
“model”: “snowboy/jarvis.pmdl”
},
“system”: “snowboy”
}
}

profile.json (not working):

{
“microphone”: {
“arecord”: {
“device”: “hw:CARD=Dummy,DEV=0”
},
“system”: “arecord”
},
“mqtt”: {
“enabled”: true,
“host”: “192.168.178.19”
},
“sounds”: {
“error”: “${RHASSPY_BASE_DIR}/sounds/wav/error.wav”,
“recorded”: “${RHASSPY_BASE_DIR}/sounds/end_of_input.wav”,
“wake”: “${RHASSPY_BASE_DIR}/sounds/start_of_input.wav”
},
“speech_to_text”: {
“system”: “kaldi”
},
“text_to_speech”: {
“system”: “picotts”
},
“wake”: {
“snowboy”: {
“model”: “snowboy/jarvis.umdl”
},
“system”: “snowboy”
}
}

Does someone has an idea how to fix this issue?

Thank you :slight_smile:

2

I overlooked the reference to https://github.com/kitt-ai/snowboy#pretrained-universal-models
For the public model jarvis.umdl you have to

  1. Set the sentitvity to “0.8,0.80”
  2. Check the box apply frontend

With those options it works :slight_smile: