VITS:Conditional Variational Autoencoder with Adversarial Learning forEnd-to-End Text-to-Speech——TTS
笔记地址:https://flowus.cn/share/4c8c251b-cb8e-4f21-aa9e-139c1c3cf883【FlowUs息流】Vits论文地址:proceedings.mlr.pressAbstract与传统的two-stageTTS(即文字→mel频谱→声音)相比,是一种parallelend-to-endTTS,提升了效率且声音自然。其它parallel方法主要存在音质