Update README.md

README.md

@@ -124,18 +124,19 @@ model, tokenizer = load_model(
 ```
 Inference:
 ```python
-def generate_response (model, tokenizer,
+def generate_response (model, tokenizer, 
                       text_input="What is the best biomaterial for superior strength?",
                       num_return_sequences = 1,
                       temperature = 0.75,  
                       max_new_tokens = 127,
                       num_beams = 1,
                       top_k = 50,
-                      top_p = 0.9, repetition_penalty=1.,
+                      top_p = 0.9,
+                      repetition_penalty=1.,
                       eos_token_id=2, 
                       add_special_tokens=True,  
                       ):
-    inputs = tokenizer(text_input,  
+    inputs = tokenizer(text_input,  add_special_tokens=add_special_tokens)
     with torch.no_grad():
           outputs = model.generate(input_ids = inputs["input_ids"],
                                     attention_mask = inputs["attention_mask"] ,