PPOCRLabel出现段错误如何debug?

[pallashadow]

操作系统：ubuntu 16.04
PaddleOCR版本：release/2.0

1、self.ocr改用GPU模式会崩溃
2、默认模型检测框过多有概率会崩溃
3、custom model用ppocr的大模型有概率会崩溃

以上的情况报错信息完全一致，没有细节：

[2021/03/09 20:03:37] root INFO: dt_boxes num : 7, elapse : 0.3803527355194092
[2021/03/09 20:03:37] root INFO: cls num : 7, elapse : 0.033449411392211914
[2021/03/09 20:03:38] root INFO: rec_res num : 7, elapse : 0.3105964660644531
QObject::killTimer: Timers cannot be stopped from another thread
QObject::~QObject: Timers cannot be stopped from another thread
[2021/03/09 20:03:38] root INFO: dt_boxes num : 5, elapse : 0.35992956161499023
[2021/03/09 20:03:38] root INFO: cls num : 5, elapse : 0.03960466384887695
[2021/03/09 20:03:38] root INFO: rec_res num : 5, elapse : 0.09412026405334473
[2021/03/09 20:03:39] root INFO: dt_boxes num : 8, elapse : 0.25769543647766113
[2021/03/09 20:03:39] root INFO: cls num : 8, elapse : 0.05037069320678711
[2021/03/09 20:03:39] root INFO: rec_res num : 8, elapse : 0.5185930728912354
[2021/03/09 20:03:40] root INFO: dt_boxes num : 26, elapse : 0.287186861038208
[2021/03/09 20:03:40] root INFO: cls num : 26, elapse : 0.11723208427429199

---

C++ Traceback (most recent call last):

0 paddle::framework::SignalHandle(char const*, int)
1 paddle::platform::GetCurrentTraceBackStringabi:cxx11

---

Error Message Summary:

FatalError: Segmentation fault is detected by the operating system.
[TimeInfo: *** Aborted at 1615291420 (unix time) try "date -d @1615291420" if you are using GNU date ***]
[SignalInfo: *** SIGSEGV (@0x6daf0fe41c91) received by PID 5216 (TID 0x7f520412a700) from PID 266607761 ***]

Segmentation fault (core dumped)

[pallashadow]

paddle.utils.run_check()
Running verify PaddlePaddle program ...
W0310 11:31:18.602231 18585 device_context.cc:362] Please NOTE: device: 0, GPU Compute Capability: 6.1, Driver API Version: 11.2, Runtime API Version: 10.2
W0310 11:31:18.604677 18585 device_context.cc:372] device: 0, cuDNN Version: 7.6.
PaddlePaddle works well on 1 GPU.
PaddlePaddle works well on 1 GPUs.
PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now.

[Evezerest]

是执行哪一步的时候出错了呢，有可能是内存不够了

[pallashadow]

是执行哪一步的时候出错了呢，有可能是内存不够了

内存和显存都还有很多
对批量图片进行auto recognition的时候。在ocr检测或识别的任意阶段都可能发生。

[Evezerest]

了解了，如果是对单个检测框重新识别会出现问题吗？以及paddle版本是多少呢

[dengrui-64]

我出现了类似的问题，跑的是自己训练的paddle-ocr模型。

背景：docker环境，用paddle检测器检测出框后，在原图上画出框，用cv2.imwrite保存图片。
原因：路径中有中文，我用的docker不支持中文，cv2.imwrite出错，莫名被paddle捕捉了错误，实际两个代码不相干。
解决：让容器支持中文环境， 运行export LC_ALL=C.UTF-8

[paddle-bot-old]

Since you haven't replied for more than 3 months, we have closed this issue/pr.
If the problem is not solved or there is a follow-up one, please reopen it at any time and we will continue to follow up.
It is recommended to pull and try the latest code first.
由于您超过三个月未回复，我们将关闭这个issue/pr。
若问题未解决或有后续问题，请随时重新打开（建议先拉取最新代码进行尝试），我们会继续跟进。