VGGish and Google Audioset

I'd like to adapt the code used for VGGish transfer learning to detect the following

features from wav files using google audioset. The training files I've downloaded, but am

new to matlab and am struggling with how to get the data into the formats needed.

Here is the link to the google site that has the classifications i want to detect:

https://research.google.com/audioset/ontology/singing_1.html

[Timothy Smith]

Google uses a tensorflow record file to save the data and can't determine how to convert this to wav files.

Answer 1

[Timothy Smith]

The solution was already provided with the Yamnet functionality in the toolbox.