Update app.py

app.py

@@ -9,6 +9,7 @@ app = Flask(__name__)
 model_id = "deepseek-ai/DeepSeek-V3-0324"
 tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True, torch_dtype=torch.float16, device_map="auto")
+model.to('cpu')
 
 @app.route('/')
 def serve_index():