diff --git a/qurator/sbb_ner/webapp/app.py b/qurator/sbb_ner/webapp/app.py
index 8820a69..ea5f4c4 100644
--- a/qurator/sbb_ner/webapp/app.py
+++ b/qurator/sbb_ner/webapp/app.py
@@ -220,6 +220,9 @@ def ner(model_id=None):
 
         for pos, (token, word_pred) in enumerate(zip(tokens, word_predictions)):
 
+            if word_pred == '[SEP]':
+                word_pred = 'O'
+
             if not token.startswith('##') and token != '[UNK]':
                 if len(word) > 0:
                     output_sentence.append({'word': word, 'prediction': last_prediction})
@@ -244,9 +247,6 @@ def ner(model_id=None):
             if not token.startswith('##') and word_pred == 'X':
                 word_pred = 'O'
 
-            if word_pred == '[SEP]':
-                word_pred = 'O'
-
             token = token[2:] if token.startswith('##') else token
 
             word += token
@@ -267,6 +267,7 @@ def ner(model_id=None):
             logger.warning('Input and output different!!! \n\n\nInput: {}\n\nOutput: {}\n'.
                            format("".join(input_sentence).replace(" ", ""),
                                   "".join([pred['word'] for pred in output_sentence])))
+    torch.cuda.empty_cache()
 
     return jsonify(output)