ML for Audio Study Group - pyctcdecode (Jan 18)

I was wondering how the hotword boosting is implemented. Is it simply changing the probabilities of the language model by a factor of x or is it something fancy?

2 Likes