Update app.py

app.py

@@ -22,7 +22,7 @@ if torch.cuda.is_available():
     model_id = "appvoid/palmer-003"
     model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=False)
     tokenizer = AutoTokenizer.from_pretrained(model_id)
-    tokenizer.use_default_system_prompt = False
+    tokenizer.use_default_system_prompt = True
 
 
 @spaces.GPU