Matt

I believe so! I'm working on documenting the pitfalls of `model.save()` with functional models, so hopefully we can avoid issues like this in future.

We also have a [KerasMetricCallback](https://huggingface.co/docs/transformers/main_classes/keras_callbacks#transformers.KerasMetricCallback) in `transformers` - but it might make sense to deprecate that in favour of a class in `evaluate`! I definitely think having an easy way...

I suspect that the layer is never really necessary when you're designing a model from scratch - you could always just use normal pooling layers and just choose the strides...

That said, it's totally okay if you want to leave it out for now - I linked it in the gist above, so feel free to just close this for...

Yes, that's correct. However, it doesn't exactly have a 'stride' in the usual sense. It basically splats (potentially overlapping) pooling windows all across the input so as to get the...

Sure, the API is just: ``` layer = AdaptiveAvgPool2D(output_dims=(128, 128)) # Can also support NCHW, but we use NHWC here inputs = tf.ones((8, 192, 192, 3), dtype=tf.float32) outputs = layer(inputs)...

In Torch it might (though I don't think this is a common/intended use case), but because I implemented it using normal pooling layers in TF I don't think it would...

Yeah, I'd say it's mostly (entirely?) useful for PyTorch model compatibility, so I get that it might feel out of place. But still, let us know if you want it...

Hi @innat I could, yes! If you look at the gist I linked above, the `pseudo_1d_pool` function is basically just a 1D `AdaptivePool`, so that would be very easy to...

Hi @dimitreOliveira, that sounds like a great feature, and we'd be happy to accept that PR! We've been working on making our default signatures more general and usable, but this...