2

app.py

@@ -125,7 +125,8 @@ def predict(message, chat_history, thinking_history, config, section_state):
         text = core_tokenizer.apply_chat_template(
             input_message,
             tokenize=False,
-            add_generation_prompt=True
+            add_generation_prompt=True,
+            enable_thinking=False
         )
         model_inputs = core_tokenizer([text], return_tensors="pt").to(core_llm.device)
         # conduct text completion