Stating the obvious: a lot of current deep learning tricks are overfit to the validation sets of well-known benchmarks, including CIFAR10. It's nice to see this quantified. This has been a problem with ImageNet since at least 2015.https://arxiv.org/abs/1806.00451
-
-
Yes, it's more expensive, but that cost itself is a regularization factor: it will force you to try fewer ideas instead of throwing spaghetti at the wall and seeing what sticks.
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I guess by "official validation set", you mean "test set", right? Otherwise, I don't really see what you mean here.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Is iterated with shuffling similar to nested CV?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
Show additional replies, including those that may contain offensive content
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.