Bole: Efficient Tree Speculation for Hybrid-Attention Language Models 文章

ArXiv CS.CL2026-08-04PAPERen作者: Li Wang, Yi Su, Xiabao Wu, Chiran You, Yongchao Liu, Zhan Qiu, Juelu Zhang, Jiajun Zheng, Fangxin Liu, Jie Zhang, Chen Tian, Chengying Huan

Bole: Efficient Tree Speculation for Hybrid-Attention Language Models · 相关公司

暂无数据