The NetApp DataOps Toolkit is a Python library that makes it simple for developers, data scientists, DevOps engineers, and data engineers to perform various data management tasks, such as near-instantaneously provisioning, cloning, or snapshotting a data volume or JupyterLab workspace.
BSD 3-Clause "New" or "Revised" License
46
stars
12
forks
source link
Request for Multi Instance GPU support for JupyterLabs #19
Can you extend support for creating JupyterLabs with MIG based GPU slices?
Typically the gpus get exposed as nvidia.com/gpu when it is not sliced with MIG but, when sliced, it gets represented as another kind of resource - nvidia.com/mig.. and so on.
Assuming that the end user knows what MIG profile to use, can we use that as an input and pass it to the deployment manifest as a resource request?
Hi,
Can you extend support for creating JupyterLabs with MIG based GPU slices? Typically the gpus get exposed as nvidia.com/gpu when it is not sliced with MIG but, when sliced, it gets represented as another kind of resource - nvidia.com/mig.. and so on.
Assuming that the end user knows what MIG profile to use, can we use that as an input and pass it to the deployment manifest as a resource request?
Thanks Arvind.