PR #16001: Add ResNet-RS to keras.applications

Imported from GitHub PR keras-team/keras#16001

**Description**

This PR adds ResNet-RS model architecture to keras.applications. refer #15780

Revisiting ResNets: Improved Training and Scaling Strategies

ResNet-RS models are updated versions of ResNet models - [Arxiv Link](https://arxiv.org/abs/2103.07579)
The models were rewritten using Keras functional API, The model's weights are converted from [original repository](https://github.com/tensorflow/tpu/tree/acb331c8878ce5a4124d4d7687df5fe0fadcd43b/models/official/resnet/resnet_rs).

**Motiviation**

Adding ResNetRS models to keras.applications has been discussed on Tensorflow's github:
#15780
It was mentioned that this family of models would be welcomed in keras.applications, hence this contribution.

**Usage / Changes**

The models could be used similarly to other models in keras.applications

```
from keras.applications import resnet_rs
model = resnet_rs.ResNetRS50()
```
**Todos**
Right now, weights are hosting on @sebastian-sz GitHub, need to move on keras-team gcs buckets.

Copybara import of the project:

--
c223693db91473c9a71c330d4e38a751d149f93c by spatil6 <sandip.aero@gmail.com>:

KERAS application addition of  Resnet-RS model

--
a1ae8cb577675e49e2b2b4963a333ee04d467978 by spatil6 <sandip.aero@gmail.com>:

KERAS application addition of Resnet-RS model, refactored code based on comments

--
9c24fc4057303172ad977cebd626da2b7adb63d4 by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor

--
2eba9003ea33054e42191a965dd45b426c0b39c3 by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor

--
780d5bc51777e1855b7df121d86f1547a30e9f54 by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor 2

--
c503d08b2ceac24e471a0739130093b66d3d0819 by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor 3

--
f4e09d4f07b3f7a5c5c5ce37793706500882e5fc by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor 4

--
5fe8ba168dd9c0cb517cfa0df39aa447df30acbc by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor 5

--
ffc5551010d9966cd20f9b8db2d7aa61ef6192aa by spatil6 <sandip.aero@gmail.com>:

Add ResNet-RS to keras.applications - code refactor 6

PiperOrigin-RevId: 437066591

RELEASE.md

@@ -64,7 +64,11 @@
 # Major Features and Improvements
 
 *   `tf.keras`:
-
+    *   Added `tf.keras.applications.resnet_rs` models.  This includes the
+        `ResNetRS50`, `ResNetRS101`, `ResNetRS152`, `ResNetRS200`,
+        `ResNetRS270`, `ResNetRS350` and `ResNetRS420` model architectures.
+        The ResNetRS models are based on the architecture described in
+        [Revisiting ResNets: Improved Training and Scaling Strategies](https://arxiv.org/pdf/2103.07579.pdf)
     *   Added `tf.keras.optimizers.experimental.Optimizer`. The reworked
         optimizer gives more control over different phases of optimizer calls,
         and is easier to customize. We provide Adam, SGD, Adadelta, AdaGrad and