Just train a new wakeword
This skill is now one of the most complex skills I have ever done and not everything has been tested and works perfectly. I appreciate any feedback
wakeword is only a few times spoken and bad detection sorted out. you could not do that through speech. i have thought about it and make a skill. This skill should make it easier for everyone to use their own precise wake words and make it possible for everyone. you can also turn on a small webserver to check your audiofiles at http://mycroft_ip:8082
- "I want to call you christopher"
- "install wakeword source"
- "You still do not understand me correctly"
- "I want to create my own model to Christopher"
- "I want to upload my wakeword Christopher"
- "I train 12 wacke word Christopher"
- "I train 12 no wacke word Christopher"
gras64
You say mycroft "I want to call you Christopher" and mycroft looks for a precise configuratios file. if under ./mycroft/precise and and in the skill folder no file is present starts learning. if there is a file, the configuration will be adjusted accordingly. You can also upgrade or upload your model to your repo and than to https://github.com/MycroftAI/Precise-Community-Data.git
if you use Public Domain Sounds Backup it will take more then 1 GB space on disk for installation. use it cairfuly!!! in the latest version i have set up a validator for the recording files. this can also be deactivated. you can also set up the number of samples to be recorded. you can also turn on a small webserver to check your audiofiles at http://mycroft_ip:8082
- use of all operating systems an configurations
- debug
- upload to https://github.com/MycroftAI/Precise-Community-Data.git
Configuration Productivity
#Wake word