reduced num_beams for beam search from 5 to 2

app.py

@@ -52,7 +52,7 @@ def generate_model_response(message: str) -> str:
     # Include **generate_kwargs to include the user-defined options
     output = model.generate(**inputs, 
                             max_new_tokens=3000, 
-                            num_beams=5,
+                            num_beams=2,
                             no_repeat_ngram_size=4,
                             early_stopping=True,
                             do_sample=True