环境：

8卡Ascend910b

npu驱动版本：23.0.3 

cann版本：8.0.0

python3.10

torch版本：2.1.0

torch_npu版本：6.0.rc3

LLM：从魔塔社区下载的Qwen2.5-7B、Qwen2-7B模型

使用ATB Models对Qwen2.5-7B、Qwen2-7B模型进行推理时，输入的命令为

torchrun --nproc_per_node 8 --master_port 20030 -m examples.run_pa --model_path /data/Qwen2-7B-Instruct/ --input_texts "What's deep learning?" "Hello World." --max_batch_size 2 

均报以下错误：

Traceback (most recent call last): 

  File "<frozen runpy>", line 198, in _run_module_as_main 

  File "<frozen runpy>", line 88, in _run_code 

  File "/usr/local/Ascend/llm_model/examples/run_pa.py", line 476, in <module> 

    pa_runner.warm_up() 

  File "/usr/local/Ascend/llm_model/examples/run_pa.py", line 204, in warm_up 

    self.model.forward( 

  File "/usr/local/Ascend/llm_model/atb_llm/runner/model_runner.py", line 157, in forward 

    return self.model.forward(**kwargs) 

           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 

  File "/usr/local/Ascend/llm_model/atb_llm/models/base/flash_causal_lm.py", line 380, in forward 

    self.init_ascend_weight() 

  File "/usr/local/Ascend/llm_model/atb_llm/models/qwen2/flash_causal_qwen2.py", line 182, in init_ascend_weight 

    self.acl_encoder_operation.set_param(json.dumps({**encoder_param})) 

RuntimeError: BuildModelGraph error,check ATB_LOG, ASDOPS_LOG