Evan Lezar

@jmpolom @jonorthwash you should be able to install the existing Debian packages on Debian stable. One thing to note is that if you are using cgroupv2 you will have to...

@DSRajesh could you give more details on image that you are using? Does `libcudnn_ops_infer.so.8` exist in the image?

@DSRajesh should the `LD_LIBRARY_PATH` not be: `LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-11.0/targets/x86_64-linux/lib/` Also, I note that the dockerfile uses `nvidia/cuda:11.0-base` as the base image. It is my understanding that the images ending in `-base` are...

The GPG keys for the CUDA download repositories were recently rotated which also affects container images using these keys. See https://forums.developer.nvidia.com/t/notice-cuda-linux-repository-key-rotation/212772 Which base image does your container use? If you're...

This is also being discussed in #1631 and #1632

@Mihawk2020 does this happen for all images or a specific one? If you start a container from the image without GPU support, does the `/usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1` file already exist?

This seems to indicate that the file `/usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1` already exists in the image. You could confirm this by running: ```bash sudo docker run --rm -ti ls -al /usr/lib/x86_64-linux-gnu/libnv\* ``` How...

Hi @startfromzero0 the CUDA driver must be available on the host, but the CUDA toolkit can be installed in a container. Alternatively one of the `nvidia/cuda` docker images can be...

No, the driver is only installed on the host. The NVIDIA Container Toolkit ensures that the required driver libraries are made available to the container.

@zhy1503 are you behind a firewall or do you need a proxy to access the internet? Note that there may be some redirects that are occurring behind the scenes.