Fix EditDistance crash with dense tensor inputs

[buildwithsuhana]

Description of the change

This PR fixes a bug where EditDistance(normalize=True) would crash when provided with a dense 2D Tensor as input. The crash occurred because the metric attempted to access the flat_values attribute, which is only available on RaggedTensor.

I have replaced the y_true.flat_values access with tf.size(y_true), which works correctly for both dense and ragged tensors in TensorFlow. A new test case has been added to edit_distance_test.py to cover this scenario.

Colab Notebook

https://colab.research.google.com/drive/1qkcPI-JIlDcaHKVTO6yx8u-gbAjUp_Hs?usp=sharing

Checklist

• I have added all the necessary unit tests for my change.
• I have verified that my change does not break existing code and works with all backends (TensorFlow, JAX, and PyTorch).
• My PR is based on the latest changes of the main branch (if unsure, rebase the code).
• I have followed the Keras Hub Model contribution guidelines in making these changes.
• I have followed the Keras Hub API design guidelines in making these changes.
• I have signed the Contributor License Agreement.

[gemini-code-assist]

Code Review

This pull request modifies the EditDistance metric to support dense tensor inputs and adds a corresponding unit test. However, the reviewer noted that using tf.size(y_true) directly will cause a crash when y_true is a tf.RaggedTensor, which is the primary input format. A code suggestion was provided to robustly handle both dense and ragged tensors.