Use conditional keras_nlp imports

[smitlg]

What does this PR do?

Uses conditional import for keras_nlp to avoid hard dependency on keras_nlp.

On the current master branch:

/keras-cv$ pip install .
/keras-cv$ python
Python 3.11.0 | packaged by conda-forge | (main, Jan 14 2023, 12:27:40) [GCC 11.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import keras_cv
...
  File "/repos/keras-cv/keras_cv/models/feature_extractor/clip/clip_processor.py", line 14, in <module>
    from keras_nlp.layers import StartEndPacker
ModuleNotFoundError: No module named 'keras_nlp'

This will require users to forcefully install keras_nlp, even when they don't want to use the part of the code requiring keras_nlp.

On this branch:

>>> import keras_cv
>>> keras_cv.__file__
'/repos/keras-cv/keras_cv/__init__.py'

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue? Please add a link
  to it if that's the case.
• Did you write any new necessary tests?
• If this adds a new model, can you run a few training steps on TPU in Colab to ensure that no XLA incompatible OP are used?

Who can review?

cc @divyashreepathihalli

[innat]

To fix the code format, make sure to run

shell/clean.sh # if keras-cv is built locally
shell/format.sh
shell/lint.sh

[smitlg]

The same error is thrown because of tensorflow_text.

[/usr/local/lib/python3.10/dist-packages/keras_cv/models/feature_extractor/clip/clip_tokenizer.py](https://localhost:8080/#) in <module>
     14 import regex as re
     15 import tensorflow as tf
---> 16 import tensorflow_text as tf_text
     17 
     18 try:

ModuleNotFoundError: No module named 'tensorflow_text'

I will update the PR to add the fix for this too.