How to pass data across pipelines in Kubeflow? For example, the output dataset from one pipeline needs to be used by another pipeline.

[Leoauro]

How to pass data across pipelines in Kubeflow? For example, the output dataset from one pipeline needs to be used by another pipeline.

[zazulam]

Hey @Leoauro, are you referring to 2 separate pipeline runs? If so, it might just be best to have those pipelines act as subpipelines inside another one.

[pritamdodeja]

Isn't this what ml metadata is supposed to help with? I've seen similar things in other frameworks where you import a particular "node", which gives you access to its outputs, which you can wire to downstream inputs.