update the query text

app.py

@@ -14,7 +14,7 @@ model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat1
 model.generation_config.pad_token_id = tokenizer.pad_token_id
 
 inner = st.text_area('enter some input!')
-text = '<|im_start|><|im_end|>user\n'+inner+'<|im_end|>\n<|im_start|>assistant\n'
+text = '<|im_start|>user\n'+inner+'<|im_end|>\n<|im_start|>assistant\n'
 
 if inner:
     inputs = tokenizer(text, return_tensors="pt")