Validation Audio - Structure of audio_val / individualAudio

[mjedmonds]

Hi,

The audio_val in tensorboard seems to contain segments of audio; what is the structure of this output? In my output, it seems to have a "burn in" before sounding similar to the input - is that expected? At what interval does it switch between validation segments?

Trying to understand the nature and structure of this output; any information is appreciated!

Thanks