How do I push a whole dataloader through a LightningModule?

[wittenator]

I migrated my model without much problems to Pytorch Lightning. Abstracting the train/eval/test loops makes sense and shortens my code significantly. Now I am wondering though how I can push a whole dataloader through my model without writing the loops myself and transferring the batches to the corresponding devices. Writing that manually doesn't seem to be in the vein of Pytorch Lightning. Is there a stylistically correct way to do that?