Add application file

app.py

@@ -118,7 +118,7 @@ def evaluate(
         )
     s = generation_output.sequences[0]
     output = tokenizer.decode(s)
-    return output.split("### Response:")[1].strip()
+    return output.split("### Response:")[1].strip().split("</s>")[0]
 
 
 g = gr.Interface(