Kedro for Deeplearning (Torch or Ternsorflow)

[vinialbert]

I'd like to understand if you have good examples of usage of Kedro for building Neural Networks pipelines?
Also, we want to understand a strategies to save intermediate training component (each epoch for example). I really can't see any solutions besides save during node processing once Kedro saves by default only in node output.