8

🎉 Our paper Factorized Diffusion Models are Natural and Zero-shot Speech Synthesizers, aka NaturalSpeech 3, got accepted by ICML 2024 as an Oral presentation!