fix

app.py

@@ -64,6 +64,7 @@ def chat_with_model(messages):
     output_text = ""
     in_think = False
     max_new_tokens = 256
+    generated_tokens = 0
 
     prompt = format_prompt(messages)
     device = torch.device("cuda")
@@ -101,7 +102,7 @@ def chat_with_model(messages):
     print(f'Step 1: {messages}')
 
     prompt_text = current_tokenizer.decode(inputs["input_ids"][0], skip_special_tokens=False)
-
+    
     for token_info in streamer:
         token_str = token_info["token"]
         token_id = token_info["token_id"]