diff --git a/.vscode/launch.json b/.vscode/launch.json
new file mode 100644
index 0000000..bbedb4b
--- /dev/null
+++ b/.vscode/launch.json
@@ -0,0 +1,25 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Python Debugger: Current File",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "${file}",
+            "console": "integratedTerminal",
+            "justMyCode": false
+        },
+        // With profiling
+        {
+            "name": "Python: Profile",
+            "type": "debugpy",
+            "request": "launch",
+            "module": "cProfile",
+            "console": "integratedTerminal",
+            "args": ["-o", "tmp.prof", "${file}"],
+        }
+    ]
+}
\ No newline at end of file
diff --git a/Demos/demos_module/iris_demo.py b/Demos/demos_module/iris_demo.py
index b17e827..36ab07a 100644
--- a/Demos/demos_module/iris_demo.py
+++ b/Demos/demos_module/iris_demo.py
@@ -48,13 +48,16 @@
 nRules = 15
 nAnts = 4
 vl = 3
-tolerance = 0.01
-fz_type_studied = fs.FUZZY_SETS.t2
+tolerance = 0.1
+fz_type_studied = fs.FUZZY_SETS.gt2
 
 # Import some data to play with
 iris = datasets.load_iris()
 X = pd.DataFrame(iris.data, columns=iris.feature_names)
 y = iris.target
+# Convert the numeric targets to class names
+# y = [iris.target_names[i] for i in y]
+class_names = list(iris.target_names)
 
 # Compute the fuzzy partitions using 3 quartiles
 precomputed_partitions = utils.construct_partitions(X, fz_type_studied)
@@ -63,11 +66,11 @@
 X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)
 
 # We create a FRBC with the precomputed partitions and the specified fuzzy set type, 
-fl_classifier = GA.BaseFuzzyRulesClassifier(nRules=nRules, linguistic_variables=precomputed_partitions, nAnts=nAnts, 
-                                            n_linguist_variables=vl, fuzzy_type=fz_type_studied, verbose=False, tolerance=tolerance, runner=runner)
-# fl_classifier.customized_loss(utils.mcc_loss)
+fl_classifier = GA.BaseFuzzyRulesClassifier(nRules=nRules, linguistic_variables=precomputed_partitions, nAnts=nAnts, class_names=class_names,
+                                            n_linguist_variables=vl, fuzzy_type=fz_type_studied, verbose=True, tolerance=tolerance, runner=runner)
+fl_classifier.customized_loss(utils.mcc_loss)
 fl_classifier.fit(X_train, y_train, n_gen=n_gen, pop_size=n_pop, checkpoints=0, random_state=0)
-print(vis_rules.rules_to_latex(fl_classifier.rule_base))
+# print(vis_rules.rules_to_latex(fl_classifier.rule_base))
 str_rules = eval_tools.eval_fuzzy_model(fl_classifier, X_train, y_train, X_test, y_test, 
                         plot_rules=True, print_rules=True, plot_partitions=True, return_rules=True)
 
diff --git a/ex_fuzzy/ex_fuzzy/rules.py b/ex_fuzzy/ex_fuzzy/rules.py
index 404f504..c0ea1dd 100644
--- a/ex_fuzzy/ex_fuzzy/rules.py
+++ b/ex_fuzzy/ex_fuzzy/rules.py
@@ -341,6 +341,9 @@ def compute_rule_antecedent_memberships(self, x: np.array, scaled=False, anteced
                 membership = np.zeros((x.shape[0], len(rule_antecedents)))
             elif self.fuzzy_type() == fs.FUZZY_SETS.t2:
                 membership = np.zeros((x.shape[0], len(rule_antecedents), 2))
+            elif self.fuzzy_type() == fs.FUZZY_SETS.gt2:
+                membership = np.zeros(
+                    (x.shape[0], len(rule_antecedents), len(self.alpha_cuts), 2))
 
             
             for ix, vl in enumerate(rule_antecedents):
@@ -737,23 +740,26 @@ def fuzzy_type(self) -> fs.FUZZY_SETS:
         return fs.FUZZY_SETS.gt2
 
 
-    def compute_rule_antecedent_memberships(self, x: np.array, scaled=True) -> np.array:
+    def compute_rule_antecedent_memberships(self, x: np.array, scaled=True, antecedents_memberships=None) -> np.array:
         '''
         Computes the membership for the antecedents performing the alpha_cut reduction.
 
         :param x: array with the values of the inputs.
         :param scaled: if True, the memberships are scaled to sum 1 in each sample.
+        :param antecedents_memberships: precomputed antecedent memberships. Not supported for GT2.
+        :return: array with the memberships of the antecedents for each sample.
         '''
         antecedent_membership = super().compute_rule_antecedent_memberships(x, scaled)
         return self._alpha_reduction(antecedent_membership)
 
 
-    def alpha_compute_rule_antecedent_memberships(self, x: np.array, scaled=True) -> np.array:
+    def alpha_compute_rule_antecedent_memberships(self, x: np.array, scaled=True, antecedents_memberships=None) -> np.array:
         '''
         Computes the membership for the antecedents for all the alpha cuts.
 
         :param x: array with the values of the inputs.
         :param scaled: if True, the memberships are scaled to sum 1 in each sample.
+        :param antecedents_memberships: precomputed antecedent memberships. Not supported for GT2.
         :return: array with the memberships of the antecedents for each sample.
         '''
         return super().compute_rule_antecedent_memberships(x, scaled)