Update app.py

app.py

@@ -253,7 +253,11 @@ def generate_qwen3_gguf(prompt: str) -> (str, str):
     messages = [
         {"role": "user", "content": prompt}
     ]
-    response = qwen3_gguf_llm.create_chat_completion(messages=messages)
+    # Set max_tokens or max_new_tokens to keep total tokens <= 512
+    response = qwen3_gguf_llm.create_chat_completion(
+        messages=messages,
+        max_tokens=512  # or smaller, adjust to fit your use case
+    )
     generated_text = response['choices'][0]['message']['content']
 
     if "</think>" in generated_text:
@@ -264,6 +268,7 @@ def generate_qwen3_gguf(prompt: str) -> (str, str):
 
 
 
+
 @app.post("/generate/{model_name}", response_model=GenerateResponse)
 async def generate(
     request: PromptRequest,