PaddleNLP [Question]: 使用UIE微调之后，在cpu上的结果和gpu的结果不一致

请提出你的问题

GPU环境：Python3.7.13 aiohttp==3.8.5 aiosignal==1.3.1 annotated-types==0.5.0 anyio==3.7.1 astor==0.8.1 async-timeout==4.0.3 asynctest==0.13.0 attrs==23.1.0 Babel==2.12.1 bce-python-sdk==0.8.90 beautifulsoup4==4.12.2 charset-normalizer==3.2.0 click==8.1.7 colorama==0.4.6 colorlog==6.7.0 concurrent-log-handler==0.9.24 cssselect==1.2.0 cycler==0.11.0 datasets==2.13.1 decorator==5.1.1 dill==0.3.4 dnspython==2.3.0 et-xmlfile==1.1.0 exceptiongroup==1.1.3 fastapi==0.103.0 filelock==3.12.2 Flask==2.2.5 flask-babel==3.1.0 fonttools==4.38.0 frozenlist==1.3.3 fsspec==2023.1.0 future==0.18.3 h11==0.14.0 html2text==2020.1.16 html5lib==1.1 httpcore==0.17.3 httpx==0.24.1 huggingface-hub==0.16.4 idna==3.4 importlib-metadata==6.7.0 itsdangerous==2.1.2 jieba==0.42.1 Jinja2==3.1.2 joblib==1.3.2 kiwisolver==1.4.5 lxml==4.9.3 markdown-it-py==2.2.0 MarkupSafe==2.1.3 matplotlib==3.5.3 mdurl==0.1.2 multidict==6.0.4 multiprocess==0.70.12.2 numpy==1.21.6 onnx==1.14.1 openpyxl==3.1.2 opt-einsum==3.3.0 packaging==23.1 paddle-bfloat==0.1.7 paddle2onnx==1.0.9 paddlefsl==1.1.0 paddlenlp==2.6.0 paddlepaddle==2.5.0 paddlepaddle-gpu==2.5.0 pandas==1.3.5 Pillow==9.5.0 portalocker==2.7.0 protobuf==3.20.2 psutil==5.9.5 py==1.11.0 pyarrow==12.0.1 pycryptodome==3.18.0 pydantic==2.3.0 pydantic_core==2.6.3 Pygments==2.16.1 pymongo==4.5.0 pyparsing==3.1.1 pyquery==2.0.0 python-dateutil==2.8.2 pytz==2023.3 PyYAML==6.0.1 rarfile==4.0 requests==2.31.0 retry==0.9.2 rich==13.5.2 safetensors==0.3.3 scikit-learn==1.0.2 scipy==1.7.3 sentencepiece==0.1.99 seqeval==1.2.2 six==1.16.0 sniffio==1.3.0 soupsieve==2.4.1 starlette==0.27.0 threadpoolctl==3.1.0 tqdm==4.66.1 typer==0.9.0 typing_extensions==4.7.1 urllib3==2.0.4 uvicorn==0.22.0 visualdl==2.5.3 webencodings==0.5.1 Werkzeug==2.2.3 xxhash==3.3.0 yarl==1.9.2 zipp==3.15.0 CPU环境： aiohttp==3.8.5 aiosignal==1.3.1 annotated-types==0.5.0 anyio==3.7.1 astor==0.8.1 async-timeout==4.0.3 asynctest==0.13.0 attrs==23.1.0 Babel==2.12.1 bce-python-sdk==0.8.90 beautifulsoup4==4.12.2 charset-normalizer==3.2.0 click==8.1.7 colorama==0.4.6 colorlog==6.7.0 concurrent-log-handler==0.9.24 cssselect==1.2.0 cycler==0.11.0 datasets==2.13.1 decorator==5.1.1 dill==0.3.4 dnspython==2.3.0 et-xmlfile==1.1.0 exceptiongroup==1.1.3 fastapi==0.103.0 filelock==3.12.2 Flask==2.2.5 flask-babel==3.1.0 fonttools==4.38.0 frozenlist==1.3.3 fsspec==2023.1.0 future==0.18.3 h11==0.14.0 html2text==2020.1.16 html5lib==1.1 httpcore==0.17.3 httpx==0.24.1 huggingface-hub==0.16.4 idna==3.4 importlib-metadata==6.7.0 itsdangerous==2.1.2 jieba==0.42.1 Jinja2==3.1.2 joblib==1.3.2 kiwisolver==1.4.5 lxml==4.9.3 markdown-it-py==2.2.0 MarkupSafe==2.1.3 matplotlib==3.5.3 mdurl==0.1.2 multidict==6.0.4 multiprocess==0.70.12.2 numpy==1.21.6 onnx==1.14.1 openpyxl==3.1.2 opt-einsum==3.3.0 packaging==23.1 paddle-bfloat==0.1.7 paddle2onnx==1.0.9 paddlefsl==1.1.0 paddlenlp==2.6.0 paddlepaddle==2.5.0 pandas==1.3.5 Pillow==9.5.0 portalocker==2.7.0 protobuf==3.20.2 psutil==5.9.5 py==1.11.0 pyarrow==12.0.1 pycryptodome==3.18.0 pydantic==2.3.0 pydantic_core==2.6.3 Pygments==2.16.1 pymongo==4.5.0 pyparsing==3.1.1 pyquery==2.0.0 python-dateutil==2.8.2 pytz==2023.3 PyYAML==6.0.1 rarfile==4.0 requests==2.31.0 retry==0.9.2 rich==13.5.2 safetensors==0.3.3 scikit-learn==1.0.2 scipy==1.7.3 sentencepiece==0.1.99 seqeval==1.2.2 six==1.16.0 sniffio==1.3.0 soupsieve==2.4.1 starlette==0.27.0 threadpoolctl==3.1.0 tqdm==4.66.1 typer==0.9.0 typing_extensions==4.7.1 urllib3==2.0.4 uvicorn==0.22.0 visualdl==2.5.3 webencodings==0.5.1 Werkzeug==2.2.3 xxhash==3.3.0 yarl==1.9.2 zipp==3.15.0 调用模型的代码： my_ie = Taskflow("information_extraction", schema=self.schema, task_path='./checkpoint/model_best') my_ie(text) cpu上返回的结果是[{}] 但实际上GPU环境下返回的结果是非空的