update app

app.py

@@ -9,7 +9,7 @@ app = FastAPI()
 
 # Load model & tokenizer sekali saat startup
 # MODEL_NAME = "Qwen/Qwen1.5-1.8B-Chat"
-MODEL_NAME = "Qwen/Qwen3-1.7B-FP8"
+MODEL_NAME = "Qwen/Qwen3Guard-Gen-0.6B"
 tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_NAME,
@@ -53,6 +53,7 @@ def chat(req: ChatRequest):
     )
 
     response = tokenizer.decode(
+        # outputs[0][inputs["input_ids"]:].tolist(),
         outputs[0][inputs["input_ids"].shape[1]:],
         skip_special_tokens=True
     )