reduced tokens to 150

app.py

@@ -107,7 +107,7 @@ def respond(message, history):
         with torch.no_grad():
             chat_history_ids = model.generate(
                 input_ids,
-                max_length=250,
+                max_length=150,
                 temperature=0.9,
                 top_p=0.9,
                 repetition_penalty=1.2,