Fetching the paper…

Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers · Around