Exploring the encoding of linguistic representations in the Fully-Connected Layer of generative CNNs for Speech

Open in new window