Tensorflow——[Could not load dynamic library cudart64_101.dll]解决方案

1. Could not load dynamic library 'cudart64_101.dll'; dlerror: cudart64_101.dll not found

2. Ignore above cudart dlerror if you do not have a GPU set up on your machine.

问题分析

1、没有安装cuda。

2、cuda版本错误。

解决方案

1、下载 CUDA Toolkit 10.1

下载地址：https://developer.nvidia.com/cuda-toolkit-archive

2、下载cudart64_101.dll

下载地址：https://cn.dll-files.com/cudart64_101.dll.html

手动将“cudart64_101.dll”添加到“C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin”中去。

DLL文件注册参考：https://www.cnblogs.com/andrew-address/p/12717201.html

3、重新安装tensorflow-gpu

4、安装tensorflow-cpu

conda install tensorflow-cpu

5、重命名

C:\Program Files\NVIDIA GPU Computing Toolkit\bin下cudart64_100.dll重命名为cudart64_101.dll

参考文章

https://blog.csdn.net/qq_41999081/article/details/104515513

https://blog.csdn.net/weixin_45485719/article/details/106399167

https://www.cnblogs.com/bigzhan/p/11855303.html

https://www.cnblogs.com/andrew-address/p/12717201.html

https://blog.csdn.net/u010094719/article/details/104571946/

转载于：https://blog.csdn.net/weixin_43272781/article/details/106417206