Implementing Adversarially Learned Inference

I’m trying to implement ALI for out of distribution (OOD) detection. How can I do this with the current fastai GAN module?

P.S.: Don’t need the code, just a direction which I can take.

1 Like

What direction did you take in the end? Working on a related problem and would love to see what you came up with!

I hacked a way out of it by changing input / output sequences. It’s just rough and would love to see what you built.