fix: correct metrics argument format in model.compile

Resolved a ValueError caused by passing a single metric object to the `metrics` argument of `model.compile`. Updated it to a list containing `tf.keras.metrics.BinaryAccuracy` to comply with Keras requirements.

Error message:
ValueError: Expected `metrics` argument to be a list, tuple, or dict.

python/dp_auditorium/dp_auditorium/testers/hockey_stick_tester.py

@@ -193,7 +193,7 @@ def _get_optimized_divergence_estimation_model(
             self._training_options.optimizer_learning_rate
         ),
         loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
-        metrics=tf.keras.metrics.BinaryAccuracy(threshold=0.0),
+        metrics=[tf.keras.metrics.BinaryAccuracy(threshold=0.0)],
     )
     features, labels = self._generate_inputs_to_model(
         samples_first_distribution,