search

StatCan / aaw-kubeflow-containers

Containers built to be used with Kubeflow for Data Science
Other
24 stars 21 forks source link

Ompp cleanup models #575

Closed Souheil-Yazji closed 9 months ago

Souheil-Yazji commented 9 months ago

No longer copying the sample models from the OMPP release to the buckets, this should also speed up the start-oms.sh script. If needed, those sample models can still be copied over manually.

vexingly commented 9 months ago

Any downside for a new namespace that will not have any models available? The web interface would just be empty and users will be responsible to compile or manually copy models before they can run/test? It does seem like for the UAT the users were only interested in utilizing their own models and not the sample models.

Souheil-Yazji commented 9 months ago

Any downside for a new namespace that will not have any models available? The web interface would just be empty and users will be responsible to compile or manually copy models before they can run/test? It does seem like for the UAT the users were only interested in utilizing their own models and not the sample models.

no downside as the folders will still be created, which means the Dir structure is preserved. This will improve the UX for users.

I'm ok with them not being there. But maybe it would be good to have one sample model in there by default so that when the openm ui starts it's not completely empty.

Honestly, from the perspective of the user, once they have their model in the bucket, it will never be empty.