T5 / Contrastive Learning

Text-aware and Context-aware Expressive Audiobook Speech Synthesis
A multi-modal, T5-based emotional-perception module for expressive audiobook TTS.
Jiaxuan Liu