Patch: Linux Host Docker (If need)

Why patch: This docker image is build in Windows host, somehow, have problems in linux host.

If you server is Linux, please run it in docker container.

nvidia-smi  #should see the GPUs, errors `not found libnvidia-ml.so.1` should patch.

So that you can enter the Docker container, and try this patch.

cd /usr/lib/x86_64-linux-gnu
ln -sf libnvidia-ml.so.5* libnvidia-ml.so.1  
ln -sf libnvidia-ml.so.5* libnvidia-ml.so
ln -sf libcuda.so.5* libcuda.so.1   #optional
ln -sf libcudadebugger.so.5* libcudadebugger.so.1   #optional
ldconfig

You can check if torch is working properly

python -c "import torch; print(torch.zeros(3,3).cuda())"

Patch: LILAB-pkg (If need)

Why patch: The docker is not support NVIDIA Video Decoder/Encoder. But the main LILAB-pkg use the NVIDIA Video packages. So the docker should downgrade to the CPU video package.

Whenever you update the LILAB-pkg from github, do this patch.

cp -r /root/Downloads/pipeline/LILAB-pkg-patch-for-docker/* ~/LILAB-pkg/

Host Port 8080 Already in Use

If you encounter the error "Error starting userland proxy: listen tcp 0.0.0.0:8080: bind: address already in use" when running the following command, it means that port 8080 on the host is already occupied.

docker run --rm -it -p 8080:8080 ...

You need to modify the port mapping to another port (e.g., 8081, http://localhost:8081/), for example:

docker run --rm -it -p 8081:8080 ...

Forward Remote Server Port 8080 to Local

You can skip this step if you are using Windows Docker.

Why forwarding is needed?

Typically, the client computer (e.g., IP 10.50.7.103) running the web interface and the remote server (e.g., IP 10.50.60.6) are on different machines. Due to Chrome's restrictions, although vs-code-server (http://10.50.60.6:8080) can be accessed, advanced features (file browsing, Jupyter, etc.) are not permitted.

Simple Solution

Forward the remote server's port 8080 (http://10.50.60.6:8080) to the local machine (http://localhost:8080, i.e., http://10.50.7.103:8080). Chrome browser can then access it normally. This is because Chrome has no restrictions on accessing localhost.

ssh -L 8080:10.50.60.6:8080 username@10.50.60.6

Operating Steps

Windows client computers can perform the following operations. Remember to replace the IP addresses with the actual server address and port.