Name Author Institution Language Exec Multiclass Regression Comments BSVM Chih-Wei Hsu and Chih-Jen Lin National Taiwan University C++ Win Yes Yes   Equbits Foresight Equbits LLC Equbits LLC SDK Win ??? Yes Commercial. Contact Equibits LLC for details Gini-SVM Shantanu Chakrabartty Johns Hopkins University C++ No Yes Yes Handles non positive definite kernels HeroSvm Jianxiong Dong Concordia University C++ Win Yes No Optimized for Pentium 4 LEARNSC Vojislav Kecman   Matlab p-files N/A Yes Yes Must pay for source! LIBSVM Chih-Chung Chang, Chih-Jen Lin National Taiwan University C++, Java, Python, R, MATLAB, Perl, Ruby Win/*nix Yes Yes Graphic interface available LS-SVMlab Kristiaan Pelckmans, Johan Suykens Katholieke Universiteit Leuven Matlab Win/*nix Yes Yes Comes with platform-specific MEX files Matlab SVM Toolbox S. R. Gunn University of Southampton Matlab N/A No Yes Includes a simple GUI mySVM Stefan Ruping Universitat Dortmund C++ Win/*nix No Yes   OSU Junshui Ma, Yi Zhao, and Stanley Ahalt Ohio State University Matlab N/A Yes Yes Matlab interface to LIBSVM Parallel GPDT T. Serafini, G. Zanghirati, L. Zanni Universita di Ferrara C++ No No No Designed for parallel systems pcSVM
Procoders.net C - No No   RVMs Mike Tipping MSR Cambridge Matlab - Yes Yes   SpiderSVM Jason Weston, Andre Elisseeff , Gokhan BakIr , Fabian Sinz Max Planck Institute for Biological Cybernetics Matlab N/A Yes Yes Part of the Spider machine learning library Statistical Pattern Recognition Toolbox for MATLAB Vojtech Franc and Vaclav Hlavac Czech Technical University Prague Matlab/C No Yes No Good online documentation. Everything I′ve tried has worked 字串9
well. Lots of stuff besides SVMs. SVMdark Martin Sewell University College London C Win No Yes   SvmFu Ryan Rifkin MIT C++ No   No Must be compiled with g++ SVMLight Thorsten Joachims Cornell University C Win/*nix No Yes   SVMsequel Hal Daume III University of Southern California OCaml No Yes No "Very fast and handles enormous datasets nicely" SVMtorch Ronan Collobert and Samy Bengio IDIAP C++ No ? Yes   SVM Toolbox Gavin Cawley University of East Anglia Norwich Matlab/C++ No Yes No Beta version WinSVM Martin Sewell University College London C++ Win No Yes  
广告合作:本站广告合作请联系QQ:858582 申请时备注:广告合作(否则不回)
免责声明:本站资源来自互联网收集,仅供用于学习和交流,请遵循相关法律法规,本站一切资源不代表本站立场,如有侵权、后门、不妥请联系本站删除!

RTX 5090要首发 性能要翻倍!三星展示GDDR7显存

三星在GTC上展示了专为下一代游戏GPU设计的GDDR7内存。

首次推出的GDDR7内存模块密度为16GB,每个模块容量为2GB。其速度预设为32 Gbps(PAM3),但也可以降至28 Gbps,以提高产量和初始阶段的整体性能和成本效益。

据三星表示,GDDR7内存的能效将提高20%,同时工作电压仅为1.1V,低于标准的1.2V。通过采用更新的封装材料和优化的电路设计,使得在高速运行时的发热量降低,GDDR7的热阻比GDDR6降低了70%。