DSPC:用于高效长上下文推理的双阶段渐进压缩框架
YaxinGao1,4YaoLu1,2,4ZongfeiZhang3JiaqiNie1,4ShanqingYu1,4QiXuan1,4
InstituteofCyberspaceSecurity,ZhejiangUniversityofTechnologyA*STARAmazon
BinjiangInstituteofArtificialIntelligence,ZhejiangUniversityofTe
DSPC:用于高效长上下文推理的双阶段渐进压缩框架
YaxinGao1,4YaoLu1,2,4ZongfeiZhang3JiaqiNie1,4ShanqingYu1,4QiXuan1,4
InstituteofCyberspaceSecurity,ZhejiangUniversityofTechnologyA*STARAmazon
BinjiangInstituteofArtificialIntelligence,ZhejiangUniversityofTe
文档评论(0)