Poly Loss

[LukeWood]

Short Description
https://arxiv.org/abs/2204.12511

Papers
https://arxiv.org/abs/2204.12511

Other Information

still not clear if we want this or not, it looks promising from the abstract but we should probably do some sanity checking.

[bhack]

It seems good, we have already 3 Pytorch third party implementations.
But I don't know if we want to collect all the bleeding edge as fast as possible without a minimum of "sedimentation" (E.g. See the experience we had with the optimizers proliferation tensorflow/addons#2203 (comment))

[LukeWood]

It seems good, we have already 3 Pytorch third party implementations. But I don't know if we want to collect all the bleeding edge as fast as possible without a minimum of "sedimentation" (E.g. See the experience we had with the optimizers proliferation tensorflow/addons#2203 (comment))

Yeah I agree. This issue is a good tracking issue; but contribution definitely not yet welcome in that we should wait to see how things unfold.

[ianstenbit]

This issue is not currently tied to any near-term objectives of KerasCV, so we are closing it for now. If you think this was done in error, or you are interested in adding this feature yourself, please checkout our README and re-open this issue if it meets the listed criteria.

Thank you! - The KerasCV team