diff --git a/Topyfic/train.py b/Topyfic/train.py
index 405ef4c..9f92be0 100644
--- a/Topyfic/train.py
+++ b/Topyfic/train.py
@@ -204,6 +204,8 @@ def save_train(self, name=None, save_path="", file_format='pickle'):
             :type name: str
             :param save_path: directory you want to use to save pickle file (default is saving near script)
             :type save_path: str
+            :param file_format: format of the file you want to save (option: pickle (default), HDF5)
+            :type file_format: str
         """
         if file_format not in ['pickle', 'HDF5']:
             sys.exit(f"{file_format} is not correct! It should be 'pickle' or 'HDF5'.")
@@ -226,9 +228,6 @@ def save_train(self, name=None, save_path="", file_format='pickle'):
             models = f.create_group("models")
             for i in range(len(self.top_models)):
                 model = models.create_group(str(i))
-
-                self.top_models[i].model = self.top_models[i].rLDA
-
                 model['components_'] = self.top_models[i].model.components_
                 model['exp_dirichlet_component_'] = self.top_models[i].model.exp_dirichlet_component_
                 model['n_batch_iter_'] = np.int_(self.top_models[i].model.n_batch_iter_)
diff --git a/Topyfic/utilsMakeModel.py b/Topyfic/utilsMakeModel.py
index da33d4c..0a6545a 100644
--- a/Topyfic/utilsMakeModel.py
+++ b/Topyfic/utilsMakeModel.py
@@ -528,15 +528,15 @@ def read_train(file):
             model = initialize_lda_model(components, exp_dirichlet_component, others)
 
             top_model = TopModel(name=f"{name}_{random_state}",
-                                         N=k,
-                                         gene_weights=components,
-                                         model=model)
+                                 N=k,
+                                 gene_weights=components,
+                                 model=model)
             top_models.append(top_model)
 
         train = Train(name=name,
-                              k=k,
-                              n_runs=n_runs,
-                              random_state_range=random_state_range)
+                      k=k,
+                      n_runs=n_runs,
+                      random_state_range=random_state_range)
         train.top_models = top_models
 
         f.close()