Update app.py

app.py

@@ -41,7 +41,8 @@ def create_response_original(input_str,
   input_ids = tokenizer.encode(input_str + tokenizer.eos_token, return_tensors="pt")
   output_ids = original_model.generate(**tokenizer(input_str+tokenizer.eos_token,return_tensors="pt",max_length=200),
                                       num_beams=num_beams, 
-                             num_return_sequences=num_return_sequences)                                 
+                             num_return_sequences=num_return_sequences,
+                                      do_sample=do_sample)                                 
   outputs = []                                 
   for output_id in output_ids:
     output = tokenizer.decode(output_id, skip_special_tokens=True)