Error in exec nvidia-smi after executing docker run --gpus all

Hey guys,
The OS is centOS7.9 and when I executed nvidia-smi on system,it’s good.everything is fine.
Then I executed " docker run -it --gpus all /bin/bash" and executed nvidia-smi on it.

Error occurred:NVIDIA-SMI couldn’t find libnvidia-ml.so library in your system. Please make sure that the NVIDIA Display Driver is properly installed and present in your system.

Please also try adding directory that contains libnvidia-ml.so to your system PATH.
I had reinstalled cuda times and change the torch version in images but still like this, and torch.cuda.is_available() still returned False.

So maybe u guys had any solution?
Thanks!

Have you tried to search for the error message? I found this: NVIDIA-SMI couldn't find libnvidia-ml.so library in your system · Issue #1163 · NVIDIA/nvidia-docker · GitHub

I searched for “NVIDIA-SMI couldn’t find libnvidia-ml.so library in your system

Unfortunately I don’t have NVIDIA on my Linux machines to try