PP-OCRv4_server_det 模型转onnx格式失败

[Longdexin]

Please fill in the information below so that we can solve the problem quickly, Thanks !

Describe the bug

/home/longdexin/python-venvs/paddle/lib/python3.10/site-packages/paddle/utils/cpp_extension/extension_utils.py:686: UserWarning: No ccache found. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://github.com/ccache/ccache/blob/master/doc/INSTALL.md
  warnings.warn(warning_message)
[Paddle2ONNX] Start to parse PaddlePaddle model...
[Paddle2ONNX] Model file path: ./inference.pdmodel
[Paddle2ONNX] Parameters file path: ./inference.pdiparams
[Paddle2ONNX] Start to parsing Paddle model...
[Paddle2ONNX] Oops, there are some operators not supported yet, including isnan_v2,
[ERROR] Due to the unsupported operators, the conversion is aborted.


--------------------------------------
C++ Traceback (most recent call last):
--------------------------------------
0   paddle2onnx::Export(char const*, char const*, char**, int*, int, bool, bool, bool, bool, bool, paddle2onnx::CustomOp*, int, char const*, char**, int*, char const*, bool*, bool, char**, int)

----------------------
Error Message Summary:
----------------------
FatalError: `Process abort signal` is detected by the operating system.
  [TimeInfo: *** Aborted at 1735875136 (unix time) try "date -d @1735875136" if you are using GNU date ***]
  [SignalInfo: *** SIGABRT (@0x3e80001cfef) received by PID 118767 (TID 0x74604816e080) from PID 118767 ***]

已中止 (核心已转储)

Informations (please complete the following information):

• Inference engine for deployment: CPU
• Why convert to onnx：PP-OCRv4_server版本模型效果很好
• Paddle2ONNX Version: 13.1
• Email/Wechat/Phone:

Screenshots

Additional context

paddle2onnx            1.3.1
paddleclas             2.6.0
paddleocr              2.9.1
paddlepaddle           3.0.0b2

[GreatV]

这应该早就修复了，使用PaddleOCR仓库导一下试试

[Longdexin]

@GreatV 好的，感谢

[Longdexin]

@GreatV 你好，git clone了PaddleOCR main分支的代码，使用pip install -e . 命令安装，发现跟之前报错一样

~/.paddlex/official_models/PP-OCRv4_server_det$ pip list |grep paddle
paddle2onnx            1.3.1
paddleclas             2.6.0
paddleocr              2.9.0.dev173+ga28196c00 /home/longdexin/PaddleOCR
paddlepaddle           3.0.0b2

[GreatV]

重新导静态图了吗

[Longdexin]

哦，没有，这个导静态图，具体怎么操作呢

[GreatV]

参考这里把你训练完成的动态图转成静态图 https://paddlepaddle.github.io/PaddleOCR/latest/ppocr/infer_deploy/python_infer.html#inference

[Longdexin]

@GreatV 非常感谢您的帮助。
看了文档，我要转的模型应该本身就是静态图模型。
我把~/.paddlex/official_models/PP-OCRv4_server_det目录的原有模型文件删除了，重新运行产线代码，自动从 https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0b2/PP-OCRv4_server_det_infer.tar 下载了新的模型文件，经测试，可以成功转为onnx格式模型，估计是官方模型更新了，有修复