diff --git a/dice_ml/explainer_interfaces/dice_genetic.py b/dice_ml/explainer_interfaces/dice_genetic.py
index 0b1bc1b7..aec92790 100644
--- a/dice_ml/explainer_interfaces/dice_genetic.py
+++ b/dice_ml/explainer_interfaces/dice_genetic.py
@@ -458,7 +458,8 @@ def find_counterfactuals(self, query_instance, desired_range, desired_class,
             population_fitness = population_fitness[population_fitness[:, 1].argsort()]
 
             current_best_loss = population_fitness[0][1]
-            to_pred = np.array([population[int(tup[0])] for tup in population_fitness[:self.total_CFs]])
+            to_pred = np.array([population[int(tup[0].item()) if hasattr(tup[0], 'item') else int(tup[0])] 
+                                for tup in population_fitness[:self.total_CFs]])
 
             if self.total_CFs > 0:
                 if self.model.model_type == ModelTypes.Classifier:
@@ -468,7 +469,8 @@ def find_counterfactuals(self, query_instance, desired_range, desired_class,
 
             # self.total_CFS of the next generation obtained from the fittest members of current generation
             top_members = self.total_CFs
-            new_generation_1 = np.array([population[int(tup[0])] for tup in population_fitness[:top_members]])
+            new_generation_1 = np.array([population[int(tup[0].item()) if hasattr(tup[0], 'item') else int(tup[0])] 
+                                         for tup in population_fitness[:top_members]])
 
             # rest of the next generation obtained from top 50% of fittest members of current generation
             rest_members = self.population_size - top_members
diff --git a/dice_ml/explainer_interfaces/explainer_base.py b/dice_ml/explainer_interfaces/explainer_base.py
index 59894f1f..f8dd9ee4 100644
--- a/dice_ml/explainer_interfaces/explainer_base.py
+++ b/dice_ml/explainer_interfaces/explainer_base.py
@@ -669,7 +669,10 @@ def misc_init(self, stopping_threshold, desired_class, desired_range, test_pred)
             self.target_cf_class = np.array(
                 [[self.infer_target_cfs_class(desired_class, test_pred, self.num_output_nodes)]],
                 dtype=np.float32)
-            desired_class = int(self.target_cf_class[0][0])
+            target_cf_class_scalar = self.target_cf_class[0][0]
+            if hasattr(target_cf_class_scalar, "item"):
+                target_cf_class_scalar = target_cf_class_scalar.item()
+            desired_class = int(target_cf_class_scalar)
             if self.target_cf_class == 0 and self.stopping_threshold > 0.5:
                 self.stopping_threshold = 0.25
             elif self.target_cf_class == 1 and self.stopping_threshold < 0.5:
@@ -692,12 +695,16 @@ def infer_target_cfs_class(self, desired_class_input, original_pred, num_output_
                 # already contains the class.
                 if hasattr(original_pred, "__len__") and len(original_pred) > 1:
                     original_pred_1 = np.argmax(original_pred)
+                    if hasattr(original_pred_1, "item"):
+                        original_pred_1 = original_pred_1.item()
                 else:
                     original_pred_1 = original_pred
                 target_class = int(1 - original_pred_1)
                 return target_class
             elif num_output_nodes == 1:  # only for pytorch DL model
                 original_pred_1 = np.round(original_pred)
+                if hasattr(original_pred_1, "item"):
+                    original_pred_1 = original_pred_1.item()
                 target_class = int(1-original_pred_1)
                 return target_class
             elif num_output_nodes > 2:
@@ -768,6 +775,8 @@ def is_cf_valid(self, model_score):
                     target_cf_class = self.target_cf_class[0]
                 elif len(self.target_cf_class.shape) == 2:
                     target_cf_class = self.target_cf_class[0][0]
+            if hasattr(target_cf_class, "item"):
+                target_cf_class = target_cf_class.item()
             target_cf_class = int(target_cf_class)
 
             if len(model_score) == 1:  # for tensorflow/pytorch models