Siva Sravana Kumar Neeli

Phase-1: #1885 - [x] Feature Pyramid Network - [x] RetinaNet Label Encoder

## Phase-2: ### TODO: - [x] Add RetinaNet Backbone Class as per KerasHub Design - [x] Add RetinaNet Task Class ## Phase -3 : ### TODO: - [ ] Train...

Hi @divyashreepathihalli, I am trying to migrate [retinanet](https://keras.io/examples/vision/retinanet/) notebook, are there any alternatives that can be used for `tf.image` functions for example `tf.image.flip_right_left`. I tried but could find any related...

@divyashreepathihalli it seems `keras_cv.layers.multi_class_non_max_suppression` does not support multi-backend as per the code [here](https://github.com/keras-team/keras-cv/blob/ce88b46eac880417096343bb8ac6b0171f0baf85/keras_cv/layers/object_detection/multi_class_non_max_suppression.py#L76C2-L76C2) and retinanet notebook uses `tf.image.combined_non_max_suppression`, it's parallel is `keras_cv.layers. multi_class_non_max_suppression`.

@VarunS1997 Seems to be `ops.take` missing `axis` argument in ROIAlign code so tried it and it seems to be working as expected. Compared same with Tensorflow Model Garden as the...

@sachinprasadhs I can take up this