1 paper
Yuancheng Wang, Zhenyu Tang, Yun Wang +9
Speech tokenizers are foundational to speech language models, yet existing approaches face two major challenges: (1) balancing trade-offs between encoding semantics for understandi…