运行tensorflow2.5训练代码之后会出现如下报错：

Loaded runtime CuDNN library: 8.0.4 but source was compiled with: 8.1.0.  CuDNN library needs to have matching major version and equal or higher minor version. If using a binary install, upgrade your CuDNN library.  If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.

更新cudnn到8.1.0即可：

conda search cudnn -c conda-forge

 conda install cudnn==8.1.0.77 -c conda-forge

更新完成之后，重新训练，没有报错了！