I was looking into some performance problems I have with Larynx and I noticed that my “Real-time factor” output is calculated wrong:
larynx: Real-time factor: 0.28 (audio=1.42 sec, infer=5.00 sec)
This was an issue but it was since fixed. So I wondered why I’m still seeing it.
Do I understand correctly that Rhasspy (the official Docker image) comes with Larynx 0.3.0 because of this line? I’m not sure how exactly the Docker image gets its Larynx, hence the question.