Update app.py

app.py

@@ -34,11 +34,11 @@ text-align: center;
 """
 
 # 加载基础模型和 LoRA 微调权重
-base_model = AutoModelForCausalLM.from_pretrained(BASE_MODEL_ID, torch_dtype=torch.float16, device_map="auto")
-tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL_ID)
+base_model = AutoModelForCausalLM.from_pretrained(BASE_MODEL_ID, torch_dtype=torch.float16, device_map="auto", use_auth_token=HF_TOKEN)
+tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL_ID, use_auth_token=HF_TOKEN)
 
 # 加载 LoRA 微调权重
-model = PeftModel.from_pretrained(base_model, LORA_MODEL_PATH)
+model = PeftModel.from_pretrained(base_model, LORA_MODEL_PATH, use_auth_token=HF_TOKEN)
 model = model.to("cuda" if torch.cuda.is_available() else "cpu")
 
 # 定义推理函数