CannyLab/tsne-cuda with cuda-10.0

t-SNE-CUDA

Barnes-Hut t-SNE

https://github.com/CannyLab/tsne-cuda/projects

做数据降维时常用到，但计算较慢，所以可用cuda加速

用源码编译时，常见错误处理：

1. 缺少MKL可到intel下载安装（略）

2. `CMake Error: The following variables are used in this project, but they are set to NOTFOUND.

如  CUDA_cublas_device_LIBRARY

可 在CMakeLists.txt

set (CUDA_cublas_device_LIBRARY /usr/local/cuda-10.0/targets/x86_64-linux/lib/libcublas.so)

libcublas.so 可用locate找到相关的版本

 

转载于:https://www.cnblogs.com/kdyi/p/10730818.html