[TFT/PyTorch] Is it possible to run the TFT implementation on Kubeflow?

[TheScorpoi]

Related to TFT/PyTorch

Question I want to know if running the TFT implementation on Kubeflow is possible and what changes I need to perform to run it in the Kubeflow environment using the Pytorch operator? Has anyone already tried this?

Given that the Pytorch operator uses docker images to train the models, and the implementation has already been made with docker images, it should work with minor changes, right?