Document return type of recording.extract_embeddings()

[thk686]

What is expected here? A list?

[Azahramirez]

It returns a dictionary with the image embedding for each 3 second frame of the spectrogram for the selected audio file