docker pull nvcr.io/nvidia/nemo:23.06
Pulling (downloading) an image might take some time.
docker run --runtime=nvidia -it --rm -v --shm-size=16g –-gpus ‘all’ -p 8888:8888 -p 6006:6006 --ulimit memlock=-1 --ulimit stack=67108864 nvcr.io/nvidia/nemo:23.06
Issuing the docker run command moves you from the host to inside the container, as shown in the following figure:
Figure 7. Output from the docker run command
Consider the following relevant information:
Figure 8. Output from the docker info command with some status information
Figure 9. Output from the docker ps command
The Docker ID is confirmed as d31c372c959c in this instance. You can usually identify which console images correspond to container consoles using the default console prompt, which includes the container ID.
Figure 10. Output from the nvidia smi command
The PyTorch release notes ( list several integrated features, including Jupyter notebooks, that this container includes. The /workspace/nemo directory in the container includes several of these features. Use commands in the container to confirm the availability of the functions that you want. There are several directories that have contents of special interest to new users such as /workspace/nemo/examples, /workspace/nemo/tutorials, and /workspace/nemo/scripts.
Figure 11. Contents of the /workspace/nemo directory