Found 1 document(s)
Page 1 of 1
-
Neural codec language models, built on transformer architecture, have revolutionized textto-speech (TTS) synthesis, excelling in voice cloning by treating it as a prefix continuation task. However, their limited context ...