1 paper
Jingbin Hu, Huakang Chen, Linhan Ma +19
Despite rapid progress in text-to-speech (TTS), open-source systems still lack truly instruction-following, fine-grained control over core speech attributes (e.g., pitch, speaking…