Can I use this work to generate/predict new audio frame in real time?

[RayShing]

Thank you for your awesome work!

I would like to ask can I use this framework to generate/predict the new audio frame based on the previous audio information and visual input in real time?