Sarvagya Malaviya

Should the layer expect the path to the mixing set (location of the fractal dataset) as an input?

I agree, I think the label smoothing issue should be fixed. Thanks for pointing it out. As for the reduction, `keras.losses.loss` already takes in a reduction argument which is handled...

Seems to run out of memory when running on `oxford_flowers102`. I didn't get time to completely investigate it but you might want to look into it once. This may help...

@LukeWood Please see if this implementation makes sense. I will further update this PR with tests and code fixes throughout the upcoming week, followed by a RetinaNet FPN implementation. The...

> This proposal includes the specific core components for building single-stage object detection models. It does not, however, include: > > Data augmentation, such as image and groundtruth box preprocessing...

> @quantumalaviya can we revive this PR as we're ready to have FPN within KerasCV? Me and @LukeWood had a discussion about how this API might make things more complicated...

> > > @quantumalaviya can we revive this PR as we're ready to have FPN within KerasCV? > > > > > > Me and @LukeWood had a discussion about...

> Where the axis=None semantic means "all the axis". > > So what do you mean with avoid reducing over any axis? As in the loss should return without reducing...

> I think that it make more sense to create another BinaryCrossentropy like BinaryCrossentropyFocal or add an extra parameter to make the mean optional in BinaryCrossentropy instead of modifying mean....