Update app.py

app.py

@@ -8,7 +8,7 @@ tokenizer = AutoTokenizer.from_pretrained(checkpoint)
 model = AutoModelForCausalLM.from_pretrained(checkpoint,torch_dtype="auto").to(device)
 
 def predict(message, history):
-    history.append({"role": "user", "assistant": message})
+    history.append({"role": "user", "content": message})
     input_text = tokenizer.apply_chat_template(history, tokenize=False)
     inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)  
     outputs = model.generate(inputs, max_new_tokens=3200, temperature=0.2, top_p=0.9, do_sample=True)