VITS2: Improving Quality and Efficiency of Single-Stage Text-to-Speech with Adversarial Learning and Architecture Design Paper • 2307.16430 • Published Jul 31, 2023 • 3