Vocoder

AbstractThis paper describes the Kyoto speech-to-speech translation system for IWSLT 2023. Our system is a combination of speech-to-text translation and text-to-speech synthesis. For the speech-to-text translation model, we used the …...