Python Kernel died

[sttpvvoyer]

Hi,

Sorry for the noob question but i'm stuck for a couple of day on that issue. I was running PaddleOCR2 sucessfully.

Upgrading to PaddleOCR3 i can run the mobile model fine.

ocr = PaddleOCR(
     text_detection_model_name="PP-OCRv5_mobile_det",
     text_recognition_model_name="PP-OCRv5_mobile_rec",
     use_doc_orientation_classify=False,
     use_doc_unwarping=False,
     use_textline_orientation=False)

As soon as i run the non-mobile i got issues with my Jupyter Lab kernel dying.

  ocr = PaddleOCR(
        det_model_dir='models/en_PP-OCRv5_server_det_infer',
        rec_model_dir='models/en_PP-OCRv5_server_rec_infer',
        use_doc_orientation_classify=False,
        use_doc_unwarping=False,
        use_textline_orientation=False
    )    

OR

ocr = PaddleOCR(use_doc_orientation_classify=False, use_doc_unwarping=False, use_textline_orientation=False)

Images are small jpg's of around 50kb with a couple 300kb. Crash is always on the 300kb images. If i OCR the 300kb individually with the non-mobile it runs fine. If i run the whole folder it seem to be the "accumulation" of predict call and that point me potentially to a memory issue?

I run this on a windows server 2019, 24gb ram and no gpu.

Any clue how to have the non-mobile version not crashing the kernel please?

[sttpvvoyer]

I've run the code in Google Colab and it make also crash the kernel. Here's the debug from Colab.

|Jun 20, 2025, 10:39:15 AM|WARNING|0.00s - Note: Debugging will proceed. Set PYDEVD_DISABLE_FILE_VALIDATION=1 to disable this validation.|
|Jun 20, 2025, 10:39:15 AM|WARNING|0.00s - to python to disable frozen modules.|
|Jun 20, 2025, 10:39:15 AM|WARNING|0.00s - make the debugger miss breakpoints. Please pass -Xfrozen_modules=off|
|Jun 20, 2025, 10:39:15 AM|WARNING|0.00s - Debugger warning: It seems that frozen modules are being used, which may|
|Jun 20, 2025, 10:39:14 AM|WARNING|WARNING:root:kernel 1142e8bc-a355-4521-bc7c-9f27a51717e5 restarted|
|Jun 20, 2025, 10:39:14 AM|INFO|KernelRestarter: restarting kernel (1/5), keep random ports|
|Jun 20, 2025, 10:39:11 AM|WARNING| [SignalInfo: *** SIGSEGV (@0x7d69b7d5f5e0) received by PID 198 (TID 0x7d6ebcf6f000) from PID 18446744072498836960 ***]|
|Jun 20, 2025, 10:39:11 AM|WARNING| [TimeInfo: *** Aborted at 1750430351 (unix time) try "date -d @1750430351" if you are using GNU date ***]|
|Jun 20, 2025, 10:39:11 AM|WARNING|FatalError: Segmentation fault is detected by the operating system.|
|Jun 20, 2025, 10:39:11 AM|WARNING|----------------------|
|Jun 20, 2025, 10:39:11 AM|WARNING|Error Message Summary:|
|Jun 20, 2025, 10:39:11 AM|WARNING|----------------------|
|Jun 20, 2025, 10:39:11 AM|WARNING|10 phi::funcs::Im2ColFunctor<(phi::funcs::ColFormat)0, phi::CPUContext, float>::operator()(phi::CPUContext const&, phi::DenseTensor const&, std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, phi::DenseTensor*, common::DataLayout)|
|Jun 20, 2025, 10:39:11 AM|WARNING|9 void phi::ConvKernelImpl<float, phi::CPUContext>(phi::CPUContext const&, phi::DenseTensor const&, phi::DenseTensor const&, std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, std::string const&, int, std::vector<int, std::allocator<int> > const&, std::string const&, phi::DenseTensor*)|
|Jun 20, 2025, 10:39:11 AM|WARNING|8 phi::KernelImpl<void (*)(phi::CPUContext const&, phi::DenseTensor const&, phi::DenseTensor const&, std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, std::string const&, std::vector<int, std::allocator<int> > const&, int, std::string const&, phi::DenseTensor*), &(void phi::ConvKernel<float, phi::CPUContext>(phi::CPUContext const&, phi::DenseTensor const&, phi::DenseTensor const&, std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, std::string const&, std::vector<int, std::allocator<int> > const&, int, std::string const&, phi::DenseTensor*))>::Compute(phi::KernelContext*)|
|Jun 20, 2025, 10:39:11 AM|WARNING|7 paddle::framework::PhiKernelInstruction::Run()|
|Jun 20, 2025, 10:39:11 AM|WARNING|6 paddle::framework::PirInterpreter::RunInstructionBase(paddle::framework::InstructionBase*)|
|Jun 20, 2025, 10:39:11 AM|WARNING|5 paddle::framework::PirInterpreter::TraceRunInstructionList(std::vector<std::unique_ptr<paddle::framework::InstructionBase, std::default_delete<paddle::framework::InstructionBase> >, std::allocator<std::unique_ptr<paddle::framework::InstructionBase, std::default_delete<paddle::framework::InstructionBase> > > > const&)|
|Jun 20, 2025, 10:39:11 AM|WARNING|4 paddle::framework::PirInterpreter::TraceRunImpl()|
|Jun 20, 2025, 10:39:11 AM|WARNING|3 paddle::framework::PirInterpreter::Run(std::vector<std::string, std::allocator<std::string > > const&, bool, bool, bool, bool)|
|Jun 20, 2025, 10:39:11 AM|WARNING|2 paddle::framework::InterpreterCore::Run(std::vector<std::string, std::allocator<std::string > > const&, bool, bool, bool, bool)|
|Jun 20, 2025, 10:39:11 AM|WARNING|1 paddle::framework::NaiveExecutor::RunInterpreterCore(std::vector<std::string, std::allocator<std::string > > const&, bool, bool)|
|Jun 20, 2025, 10:39:11 AM|WARNING|0 paddle::AnalysisPredictor::ZeroCopyRun(bool)|
|Jun 20, 2025, 10:39:11 AM|WARNING|--------------------------------------|
|Jun 20, 2025, 10:39:11 AM|WARNING|C++ Traceback (most recent call last):|
|Jun 20, 2025, 10:39:11 AM|WARNING|--------------------------------------|

[sttpvvoyer]

This image specifically (and some other too) seem to make the kernel crashing....

[sttpvvoyer]

That continues on random images. All of them with hand written character. I've manage to process some by resizing them or by "saving for web". Some i can't at all, the kernel always dying. Please, any help?