nvidia 1080Ti显卡，cuda10.2，使用paddleOCR 推理报错：

[wenxuezhang]

• 系统环境/System Environment：
• centos 7.7 ；nvidia 1080Ti显卡，cuda10.2， cudnn7.6.5.32，
• 版本号/Version：Paddle： PaddleOCR： 问题相关组件/Related components： paddle 版本：paddlepaddle-gpu 2.4.2 paddleOCR 版本：paddleocr 2.6.1.3

模型：

方向分类器

cls_model:ch_ppocr_mobile_v2.0_cls_infer

ocr检测模型:

det_model: ch_PP-OCRv3_det_infer

ocr识别模型

rec_model: ch_PP-OCRv3_rec_infer rec_char_dict_path: /ppocr_keys_v1.txt font_path:sim_fang.ttf char_type: ch

• 运行指令/Command Code： ocr = PaddleOCR(cls_model_dir=cls_model,det_model_dir=det_model, rec_model_dir=rec_model, rec_char_dict_path = rec_char_path, rec_char_type=rec_char_type,version='PP-OCRv3', vis_font_path=font_path, use_angle_cls=False,use_gpu=True)

执行：ocr.ocr(file_name, cls=True)，报错如下：

• 完整报错/Complete Error Message：
• terminate called after throwing an instance of 'phi::enforce::EnforceNotMet' what(): (External) CUDA error(3), initialization error. [Hint: Please search for the error code(3) on website (https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html#group__CUDART__TYPES_1g3f51e3575c2178246db0a94a430e0038) to get Nvidia's official solution and advice about CUDA Error.] (at /paddle/paddle/phi/backends/gpu/cuda/cuda_info.cc:172)

---

C++ Traceback (most recent call last):

0 void paddle_infer::Tensor::CopyFromCpu(float const) 1 std::basic_future<std::unique_ptr<phi::DeviceContext, std::default_delete > >::_M_get_result() const 2 std::future_base::_Deferred_state<std::thread::_Invoker<std::tuple<std::unique_ptr<phi::DeviceContext, std::default_delete > ()(phi::Place const&, bool), phi::Place, bool> >, std::unique_ptr<phi::DeviceContext, std::default_delete > >::_M_complete_async() 3 std::future_base::_State_baseV2::_M_do_set(std::function<std::unique_ptr<std::__future_base::_Result_base, std::future_base::_Result_base::_Deleter> ()>, bool) 4 std::_Function_handler<std::unique_ptr<std::__future_base::_Result_base, std::future_base::_Result_base::_Deleter> (), std::future_base::_Task_setter<std::unique_ptr<std::future_base::_Result<std::unique_ptr<phi::DeviceContext, std::default_delete > >, std::future_base::_Result_base::_Deleter>, std::thread::_Invoker<std::tuple<std::unique_ptr<phi::DeviceContext, std::default_delete > (*)(phi::Place const&, bool), phi::Place, bool> >, std::unique_ptr<phi::DeviceContext, std::default_delete > > >::_M_invoke(std::_Any_data const&) 5 std::unique_ptr<phi::DeviceContext, std::default_delete > paddle::platform::CreateDeviceContext(phi::Place const&, bool) 6 phi::GPUContext::Impl::~Impl()

---

Error Message Summary:

FatalError: Process abort signal is detected by the operating system. [TimeInfo: Aborted at 1677048461 (unix time) try "date -d @1677048461" if you are using GNU date ] [SignalInfo: SIGABRT (@0x62e1) received by PID 25313 (TID 0x7f87bca50700) from PID 25313 ]

[tink2123]

请检查 paddle 安装是否正确：

import paddle
paddle.utils.run_check()

[wenxuezhang]

import paddle paddle.utils.run_check() Running verify PaddlePaddle program ... W0222 22:25:37.441035 22594 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 6.1, Driver API Version: 10.2, Runtime API Version: 10.2 W0222 22:25:37.451232 22594 gpu_resources.cc:91] device: 0, cuDNN Version: 7.6. PaddlePaddle works well on 1 GPU. /opt/anaconda3/envs/py37ocr/lib/python3.7/site-packages/paddle/fluid/executor.py:1585: UserWarning: Standalone executor is not used for data parallel UserWarning) W0222 22:25:40.725191 22594 parallel_executor.cc:666] Cannot enable P2P access from 0 to 1 W0222 22:25:40.725250 22594 parallel_executor.cc:666] Cannot enable P2P access from 1 to 0 W0222 22:25:43.492534 22594 fuse_all_reduce_op_pass.cc:79] Find all_reduce operators: 2. To make the speed faster, some all_reduce ops are fused during training, after fusion, the number of all_reduce ops is 2. PaddlePaddle works well on 2 GPUs. PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now. 我推理的时候用的是一个卡

[wenxuezhang]

为啥关闭这个issue? 问题并没有解决。硬件检测通过说明paddleocr 安装的没问题，然后我加载模型也没问题，但是我推理的时候出错啊，我上面描述的错误是推理的错误

[wenxuezhang]

我单独以交互式方式执行是没问题的。但是通过tornado放在进程里就不好使，不知道咋回事，有类似的issue 能推荐下吗？我觉得这个问题应该不是个特别的问题

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.