Assigned label speaker identification example

[razi-tm]

@hbredin @flyingleafe
#1346
I have managed to extract embeddings using "speechbrain/spkrec-ecapa-voxceleb" model. Are you using this model for extracting the embeddings? Can you provide an example of how to extract embeddings resulting in Having the audio segments labeled based on those embeddings? (like this issue: #1383 (comment) by @Vermeille)
I think this implementation example can be really clarifying and useful for many people.

Thanks.