Get last embedding layer from wav2vec

I see there’s something called last_hidden_state how do I pull the embedding of the entire audio sequence from the Wav2Vec model (Wav2Vec2)?