update generate parameters

app.py

@@ -41,7 +41,8 @@ def chat_func(input_utterance: str, history: Optional[List[str]] = None):
                                 top_p=0.95,
                                 do_sample=True,
                                 num_beams=4,
-                                repetition_penalty=1.2)[0]
+                                repetition_penalty=1.2,
+                                no_repeat_ngram_size=4)[0]
 
     response = tokenizer.decode(output_ids, skip_special_tokens=True)