Kubeflow Local execution mode

[Saad Zaher]

Hello everyone,

We are working on a local execution mode for Kubeflow Trainer SDK to enable data scientists and ML engineers to validate their jobs locally first on docker before submitting it to the Kubeflow Trainer. 

The proposal doc can be found here: https://docs.google.com/document/d/1E-bCT831f6GE7S6TvPecVpsnzrtEyVMfF52i866aCoE/edit?tab=t.0 To be transferred to KEP shortly.

We appreciate feedback from the community. 

Thanks,

Saad!