bhack
bhack
/cc @howl-anderson
We could take a look at: https://github.com/tensorflow/addons/blob/v0.15.0/tensorflow_addons/image/dense_image_warp.py https://github.com/tensorflow/addons/blob/v0.15.0/tensorflow_addons/image/sparse_image_warp.py
> I think that the dense_image can help with the last stage of elastic deformation. Dense_image_warp needs a list of sample points, which means the transform has to be computed...
/cc @yuanliangzhe
@shkarupa-alex Are you still interested in this?
Can you check the function we have in [Tensorflow 3d](https://github.com/google-research/google-research/blob/master/tf3d/object_detection/box_utils/box_ops.py)? Probably they could be interested to have a high performance implementation of this op. It was presented one months ago:...
I am monitoring your last comment at https://github.com/google-research/google-research/issues/660
It could be interesting to have a learnable 3d nms like: https://github.com/abhi1kumar/groomed_nms/blob/main/lib/groomed_nms.py
Probably this component is orphan. Gently ping @szutenberg @markub3327 Can you check with a ticket on https://github.com/keras-team/keras if they are interested if you port this in Keras with your `exclude_from_weight_decay`...
Thank you as `AdamW` was just duplicated [there 27 days ago](https://github.com/keras-team/keras/commit/3b87e6c50be9cd8618aab5b154a72c20a6b97f9b) probably you could add a comment there also for `exclude_from_weight_decay` for `AdamW`