feat: Changes related to Stage 1 - TF backend compatibility check for Vision Transformer on small datasets

[pksX01]

This PR caters to address Stage 1: tf.keras backwards compatibility check of the Issue Keras.io examples conversion gameplan #18468 for vitsmall_ds.py example with some essential modification to make script executable error free for tensorflow backend.

Here is the git diff with original file -

diff --git a/examples/vision/vit_small_ds.py b/examples/vision/vit_small_ds.py
index 7fd77de..31be894 100644
--- a/examples/vision/vit_small_ds.py
+++ b/examples/vision/vit_small_ds.py
@@ -5,6 +5,7 @@ Date created: 2022/01/07
 Last modified: 2022/01/10
 Description: Training a ViT from scratch on smaller datasets with shifted patch tokenization and locality self-attention.
 Accelerator: GPU
+Keras 3 Conversion initiated by: [Pavan Kumar Singh](https://github.com/pksX01)
 """
 """
 ## Introduction
@@ -33,26 +34,20 @@ The main ideas are:
 This example implements the ideas of the paper. A large part of this
 example is inspired from
 [Image classification with Vision Transformer](https://keras.io/examples/vision/image_classification_with_vision_transformer/).
-
-_Note_: This example requires TensorFlow 2.6 or higher, as well as
-[TensorFlow Addons](https://www.tensorflow.org/addons), which can be
-installed using the following command:
-
-```python
-pip install -qq -U tensorflow-addons
-```
 """
 """
 ## Setup
 """

+import os
+os.environ["KERAS_BACKEND"] = "tensorflow"
+
 import math
 import numpy as np
 import tensorflow as tf
-from tensorflow import keras
-import tensorflow_addons as tfa
+import keras
 import matplotlib.pyplot as plt
-from tensorflow.keras import layers
+from keras import layers

 # Setting seed for reproducibiltiy
 SEED = 42
@@ -354,7 +349,7 @@ at a later stage.
 """


-class MultiHeadAttentionLSA(tf.keras.layers.MultiHeadAttention):
+class MultiHeadAttentionLSA(keras.layers.MultiHeadAttention):
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
         # The trainable temperature term. The initial value is
@@ -498,7 +493,7 @@ def run_experiment(model):
         warmup_steps=warmup_steps,
     )

-    optimizer = tfa.optimizers.AdamW(
+    optimizer = keras.optimizers.AdamW(
         learning_rate=LEARNING_RATE, weight_decay=WEIGHT_DECAY
     )

@@ -547,7 +542,4 @@ supplementary of the paper.

 I would like to thank [Jarvislabs.ai](https://jarvislabs.ai/) for
 generously helping with GPU credits.
-
-You can use the trained model hosted on [Hugging Face Hub](https://huggingface.co/keras-io/vit_small_ds_v2)
-and try the demo on [Hugging Face Spaces](https://huggingface.co/spaces/keras-io/vit-small-ds).
 """

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (10252a9) 63.86% compared to head (56bde42) 79.44%.
Report is 5 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master   #18884       +/-   ##
===========================================
+ Coverage   63.86%   79.44%   +15.57%     
===========================================
  Files         336      336               
  Lines       34848    34854        +6     
  Branches     6855     6851        -4     
===========================================
+ Hits        22255    27689     +5434     
+ Misses      11121     5584     -5537     
- Partials     1472     1581      +109     

Flag	Coverage Δ
keras	79.30% <ø> (+15.44%)	⬆️
keras-jax	61.04% <ø> (?)
keras-numpy	55.96% <ø> (?)
keras-tensorflow	63.21% <ø> (?)
keras-torch	63.84% <ø> (-0.03%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[fchollet]

Thanks for the PR! The code changes look good. Please reopen the PR on https://github.com/keras-team/keras-io -- we have moved example handling there. Keras 3 examples were only hosted on keras-team/keras temporarily while keras.io was still focused on Keras 2.

I have updated the instructions in the "call for contributions" pinned issue.