Update app.py

app.py

@@ -49,7 +49,7 @@ def load_models():
     
     # 2. Load text generation model
     # llm_model_name = "Qwen/Qwen3-0.6B"
-    llm_model_name = "meta-llama/Meta-Llama-3-8B"
+    llm_model_name = "Qwen/QwQ-32B"
     tokenizer = AutoTokenizer.from_pretrained(llm_model_name)
     llm_model = AutoModelForCausalLM.from_pretrained(
         llm_model_name,
@@ -145,7 +145,7 @@ def generate_response(message, chat_history, analysis_results):
         model_inputs = tokenizer([text], return_tensors="pt").to(llm_model.device)
         generated_ids = llm_model.generate(
             **model_inputs,
-            max_new_tokens=256,
+            max_new_tokens=32768,
             do_sample=True,
             temperature=0.7,
         )