Open jiaxialxj opened 2 years ago
My environment is here: Ubuntu 20.04 Kernel version: 5.8.0-43-generic Nvidia Driver version: 460.27.04 CUDA Version: 11.2.0 cuDnn version: 8.1.1 Tensorflow version: 2.6.0
Hi,
Based on our previous experience, this error could be caused by not properly loading Cudnn. Could you add "--log_level debug" in your command and check the output?
Best,
My environment is here: Ubuntu 20.04 Kernel version: 5.8.0-43-generic Nvidia Driver version: 460.27.04 CUDA Version: 11.2.0 cuDnn version: 8.1.1 Tensorflow version: 2.6.0