Fix callback compatibility for trainable_model

qiskit_machine_learning/algorithms/trainable_model.py

@@ -20,7 +20,7 @@
 
 from .objective_functions import ObjectiveFunction
 from .serializable_model import SerializableModelMixin
-from ..optimizers import Optimizer, SLSQP, OptimizerResult, Minimizer
+from ..optimizers import Optimizer, SciPyOptimizer, SLSQP, OptimizerResult, Minimizer
 from ..utils import algorithm_globals
 from ..neural_networks import NeuralNetwork
 from ..utils.loss_functions import (
@@ -269,7 +269,8 @@ def _get_objective(
 
         def objective(objective_weights):
             objective_value = function.objective(objective_weights)
-            self._callback(objective_weights, objective_value)
+            if isinstance(self._optimizer, SciPyOptimizer):
+                self._callback(objective_weights, objective_value)
             return objective_value
 
         return objective

releasenotes/notes/fix-callback-trainable_model-ca52060260a3e466.yaml

@@ -0,0 +1,5 @@
+---
+fixes:
+  - |
+      Fixed a compatibility issue for different `callback` functions when they are 
+      used in :class:`~qiskit_machine_learning.algorithms.trainable_model` based algorithms.

test/algorithms/classifiers/test_neural_network_classifier.py

@@ -521,17 +521,18 @@ def test_callback_setter(self):
 
         loss_history = []
 
-        def store_loss(_, loss):
-            loss_history.append(loss)
+        def store_loss(nfev, x_next, loss, update, is_accepted):  # pylint: disable=unused-argument
+            if is_accepted:
+                loss_history.append(loss)
 
         # use setter for the callback instead of providing in the initialize method
-        classifier.callback = store_loss
+        classifier.optimizer.callback = store_loss
 
         features = np.array([[0, 0], [1, 1]])
         labels = np.array([0, 1])
         classifier.fit(features, labels)
 
-        self.assertEqual(len(loss_history), 3)
+        self.assertEqual(len(loss_history), 1)
 
 
 if __name__ == "__main__":

test/algorithms/regressors/test_neural_network_regressor.py

@@ -239,17 +239,19 @@ def test_callback_setter(self):
 
         loss_history = []
 
-        def store_loss(_, loss):
-            loss_history.append(loss)
+        def store_loss(nfev, x_next, loss, update, is_accepted):  # pylint: disable=unused-argument
+
+            if is_accepted:
+                loss_history.append(loss)
 
         # use setter for the callback instead of providing in the initialize method
-        regressor.callback = store_loss
+        regressor.optimizer.callback = store_loss
 
         features = np.array([[0, 0], [0.1, 0.1], [0.4, 0.4], [1, 1]])
         labels = np.array([0, 0.1, 0.4, 1])
         regressor.fit(features, labels)
 
-        self.assertEqual(len(loss_history), 3)
+        self.assertEqual(len(loss_history), 1)
 
 
 if __name__ == "__main__":