Gerald Monroe comments on how 2 tell if ur input is out of distribution given only model weights

[Gerald Monroe]

You could also train an autoencoder on the training set and then autoencode the input image and measure the residual part.

[dkirmani]

Yep, but:

given only model weights

Also, the cossim-based approach should work on arbitrary submodules of a given model! Also it’s fast!