fix: tf backend execute_with_gradients when xs contains KerasVariable

ivy/functional/backends/tensorflow/gradients.py

@@ -76,6 +76,16 @@ def execute_with_gradients(
         _get_required_float_variables(xs, xs_grad_idxs)
     )
 
+    # Conversion of KerasVariable to tf.Variable within xs_required container, so they can be watched
+    if ivy.is_ivy_container(xs_required):
+        ivy.nested_map(
+            lambda x: x._value
+            if "keras.src.backend.tensorflow.core.Variable" in str(x.__class__)
+            else x,
+            xs_required,
+            include_derived=True,
+        )
+
     # Creating a tape to record operations
     with tf.GradientTape(persistent=True, watch_accessed_variables=False) as tape:
         tape.watch(xs_required)