Errors while running classification_with_tfdf from keras-io

[Humbulani1234]

The script has the following error:

Traceback (most recent call last):
  File "/home/humbulani/tensorflow-env/keras-io-master/examples/structured_data/classification_with_tfdf.py", line 290, in <module>
    gbt_model = create_gbt_model()
                ^^^^^^^^^^^^^^^^^^
  File "/home/humbulani/tensorflow-env/keras-io-master/examples/structured_data/classification_with_tfdf.py", line 282, in create_gbt_model
    gbt_model.compile(metrics=[keras.metrics.BinaryAccuracy(name="accuracy")])
  File "/home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tensorflow_decision_forests/keras/core_inference.py", line 852, in compile
    super(InferenceCoreModel, self).compile(
  File "/home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tf_keras/src/utils/traceback_utils.py", line 70, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tensorflow/python/distribute/distribute_lib.py", line 4029, in variable_created_in_scope
    return v._distribute_strategy is None  # pylint: disable=protected-access
           ^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'Variable' object has no attribute '_distribute_strategy'

I have traced the error to the following section of code from the function tf_keras.src.engine.training._validate_compile:

strategy = self.distribute_strategy
        for metric in tf.nest.flatten(metrics):
            for v in getattr(metric, "variables", []): # ERROR IS HERE
                if not strategy.extended.variable_created_in_scope(v):
                    raise ValueError(
                        f"Metric ({metric}) passed to `model.compile` was "
                        "created inside a different distribution strategy "
                        "scope than the model. All metrics must be created "
                        "in the same distribution strategy "
                        f"scope as the model (in this case {strategy}). "
                        "If you pass in a string identifier for a metric to "
                        "compile, the metric will automatically be created "
                        "in the correct distribution strategy scope."
                    )

and pdb has produced the following trace:

(Pdb) n
> /home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tf_keras/src/engine/training.py(3927)_validate_compile()
-> strategy = self.distribute_strategy
(Pdb) n
> /home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tf_keras/src/engine/training.py(3928)_validate_compile()
-> for metric in tf.nest.flatten(metrics):
(Pdb) p strategy
<tensorflow.python.distribute.distribute_lib._DefaultDistributionStrategy object at 0x7eda39f63750>
(Pdb) list
3923 	                    )
3924 	
3925 	        # Model metrics must be created in the same distribution strategy scope
3926 	        # as the model.
3927 	        strategy = self.distribute_strategy
3928 ->	        for metric in tf.nest.flatten(metrics):
3929 	            for v in getattr(metric, "variables", []):
3930 	                if not strategy.extended.variable_created_in_scope(v):
3931 	                    raise ValueError(
3932 	                        f"Metric ({metric}) passed to `model.compile` was "
3933 	                        "created inside a different distribution strategy "
(Pdb) n
> /home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tf_keras/src/engine/training.py(3929)_validate_compile()
-> for v in getattr(metric, "variables", []):
(Pdb) p metric
<BinaryAccuracy name=accuracy>
(Pdb) n
> /home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/tf_keras/src/engine/training.py(3930)_validate_compile()
-> if not strategy.extended.variable_created_in_scope(v):
(Pdb) p v
<Variable path=accuracy/total, shape=(), dtype=float32, value=0.0>
(Pdb) p v._distribute_strategy
*** AttributeError: 'Variable' object has no attribute '_distribute_strategy'
(Pdb) p metric.variables
[<Variable path=accuracy/total, shape=(), dtype=float32, value=0.0>, <Variable path=accuracy/count, shape=(), dtype=float32, value=0.0>]