原创力文档OSCAR 面向 2 比特键值缓存量化的离线谱协方差感知旋转变换算法 OSCAR Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization.pdfVIP
- 0
- 0
- 约23.05万字
- 约 35页
- 2026-09-21 发布于广东
- 举报
OSCAR:OfflineSpectralCovariance-AwareRotation
for2-bitKVCacheQuantization
6
2
0
2
yZhongzhuZhou∗,1,2DonglinZhuang∗,1,2JisenLi1,3ZiyanChen2
aShuaiwenLeonSong1BenAthiwaratkun1XiaoxiaWu†,1
M1TogetherAI2UniversityofSydney3UniversityofIllinoisUrbana-Champaign
8∗Equalcontribution.†Correspondingauthor.
1
]
G
您可能关注的文档
- nber-地缘政治分裂、主权债务和美元主导 Geopolitical Fragmentation, Sovereign Debt, and Dollar Dominance.pdf
- nber-定义创新 NewSpace 和不断变化的空间部门的案例 Defining Innovatisation The Case of NewSpace and the Changing Space Sector.pdf
- nber-多产品生产中的因素偏差与异质产出质量 Factor Bias and Heterogeneous Output Quality in Multiproduct Production.pdf
- nber-非用户外部性 Non-User Externalities.pdf
- nber-非洲的创业公司 Startups in Africa.pdf
- nber-风中的尘埃：盐湖严重干燥的健康成本 Dust in the Wind The Health Costs of Great Salt Lake Desiccation.pdf
- nber-进口入侵是否解释了美国制造业生产率增长的神秘消失 Does the Import Invasion Explain the Mysterious Disappearance of Productivity Growth in U.S. Manufacturing.pdf
- nber-具有异质性偏好和风险的最佳长期护理提供和保险 Optimal Long-Term Care Provision and Insurance with Heterogeneous Preferences and Risks.pdf
- nber-劳动力供给曲线外：大公司的零雇主规模工资效应 Off the Labor Supply Curve The Zero Employer Size Wage Effect Within Large Firms.pdf
- nber-联邦政府的自由支配支出 The Federal Government's Discretionary Spending.pdf
- oxfordeconomics -安大略省市政协会安大略省市政资本支出报告中的关税影响 TARIFF IMPACTS ON ONTARIO MUNICIPALITY CAPITAL EXPENDITURE REPORT FOR ASSOCIATION OF MUNICIPALITIES OF ONTARIO 2026.pdf
- oxfordenergy - 俄罗斯核能巨头俄罗斯国家原子能公司影响力持续攀升 Rosatom – the growing influence of Russia’s nuclear powerhouse.pdf
- oxfordenergy - 欧洲需要多少液化天然气来补充储气库存 How much LNG does Europe need to refill storage.pdf
- PhysX-Omni：面向刚体、可变形体与铰接物体的一体化可仿真物理三维生成方案 PhysX-Omni Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objects.pdf
- Primer PICKR 基于文献挖掘、用于筛选高稳定性实时荧光定量 PCR 引物的评分平台 Primer PICKR literature-mined scoring platform for robust RT–qPCR primers.pdf
- rand -打击涉芬太尼贩运问题 减少毒品供给与开展国际协作的相关政策考量 Combating Chinese Fentanyl Trafficking Policy Considerations for Supply Reduction and International Coordination.pdf
- ris -2026 年世界贸易和发展报告 World Trade and Development Report 2026.pdf
- RWA新规：新加坡金管局降低无许可区块链数字资产的银行资本金要求.pdf
文档评论（0）