use cache

app.py

@@ -47,6 +47,7 @@ def predict(message, history, system_prompt, temperature, max_tokens):
         top_p=0.95,
         temperature=0.5,
         max_new_tokens=DEFAULT_MAX_NEW_TOKENS,
+        use_cache=True,
     )
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     t.start()