paddlelite移动端端部署v3模型多线程报错

[Onmyown577]

paddlelite版本：2.11和2.10 部署模型PaddleOCRv3模型 测试系统：Android 问题描述：移动端4线程预测时paddlelit2.11和2.10都报Segmentation fault错误 threads 4 power_mode 3 具体报错： /Paddle-Lite/lite/core/device_info.cc:1118 Setup] ARM multiprocessors name: HARDWARE : MT6762V/CB M101_MT6765 /Paddle-Lite/lite/core/device_info.cc:1119 Setup] ARM multiprocessors number: 8 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 0, max freq: 2001, min freq: 2001, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 1, max freq: 2001, min freq: 2001, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 2, max freq: 2001, min freq: 2001, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 3, max freq: 2001, min freq: 2001, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 4, max freq: 1500, min freq: 1500, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 5, max freq: 1500, min freq: 1500, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 6, max freq: 1500, min freq: 1500, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1121 Setup] ARM multiprocessors ID: 7, max freq: 1500, min freq: 1500, cluster ID: 0, CPU ARCH: A53 /Paddle-Lite/lite/core/device_info.cc:1127 Setup] L1 DataCache size is: /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1129 Setup] 32 KB /Paddle-Lite/lite/core/device_info.cc:1131 Setup] L2 Cache size is: /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1133 Setup] 512 KB /Paddle-Lite/lite/core/device_info.cc:1135 Setup] L3 Cache size is: /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1137 Setup] 0 KB /Paddle-Lite/lite/core/device_info.cc:1139 Setup] Total memory: 3882740KB Segmentation fault (core dumped)

[LDOUBLEV]

单线程会报错吗？

可以在代码里std::cout 一些log，看下是在哪里出现段错误的。

用v2的模型会不会出现错误

[Onmyown577]

单线程都可以运行，在Android9上2.11的四线程可以，但是在Android11上2.11不能，2.10两个系统都会报错 V2模型还没有验证

[zhangyingbit]

@LDOUBLEV 大佬， 想问下使用paddle lite在x86上部署，目前v3 det 和 rec都已经跑通，但如果运行system的预测就会报double free的错，我打了下log， 发现是predictor_crnn->Run() 这个操作在识别第二个boxes 的文本行报错，进行了一下debug （1）system输入的图像由多行文本行换成单行文本行的图像，不报错，程序正常运行 （2）把ocr_db_crnn.cc中以下代码boxes[i]换成boxes[boxes.size() -2]，也不会报错，程序正常运行 for (int i = boxes.size() - 1; i >= 0; i--) { auto preprocess_start = std::chrono::steady_clock::now(); crop_img = GetRotateCropImage(srcimg, boxes[i]);

我推测是预测的boxes的大小不一，但是识别网络的输入没有处理好应对长度不一的输入，请教下您该如何解决呢？被困了快一周了，谢谢~