Common errors of running deep learning

1. Loaded runtime CuDNN library: 7103 (compatibility version 7100) but source was compiled with 7003

The problem is (a) you may have multiple versions of CUDNN or (b)your tensorflow version is not matching the CUDNN version.

solution: for (a), you may delete all versions of CUDNN and reinstall it; for (b), you can upgrade or downgrade your tensorflow version, such as conda install tensorflow-gpu==1.8