documentation for recognition of multiple wakewords

[pvanickova]

Hello,
can you please add documentation on how to detect multiple wakewords? Is it preferable to train a separate model for each wakeword or to train a single model supporting several wakewords?

Also do the "spot" and "test" commands support multi wakeword detection?