Matt

@jbischof The layer is generally found inside [pyramid pooling](https://arxiv.org/abs/1406.4729) modules. In :hugs: `transformers` we have several models that use it in the PyTorch implementation. All of these need a TF...

We also have lots of other uses of it in the codebase that just use it to resize down to (1, 1). In those cases, though, we can just port...

@jbischof Likely yes! Right now we've just had to implement TF AdaptivePool in our own codebase, so if there was a proper reference version in KerasCV we'd probably switch to...

@awsaf49 It totally fell off the radar, unfortunately, because of a pile of other stuff to do in `transformers`! Right now our TF efforts are focused on Keras 3.0 compatibility...

This looks like an issue in the Torch notebook rather than the TF one - @amyeroberts do you know who wrote the original torch version?

Yes, sorry for not checking in @yonigottesman! Do you have other features you want to add, or should we treat this as review-ready?

On it!

@yonigottesman while I'm reviewing can you rebase/resolve the merge conflict? It's nothing major, but it'll block us merging the PR until it's ready. (Edit: Probably better to rebase because your...

Yes, agree! It's also fine to leave that for a separate PR, and just add the mask functionality in this PR.

Got it! Ping me whenever you're ready for re-review.