![](https://img-blog.csdnimg.cn/20201014180756925.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
Debug
Mr.Q
纸上得来终觉浅,绝知此事要躬行。 ---陆游
展开
-
visual studio打包qt算子时,只生成dll没有生成lib等文件
解决不生成lib文件问题。原创 2024-06-03 15:55:42 · 323 阅读 · 0 评论 -
qt.qpa.plugin:Could not load the Qt platform plugin “xcb“
错误信息中提到“"/home/zxq/anaconda3/envs/open3d/lib/python3.9/site-packages/cv2/qt/plugins”解决办法:删除上面目录中的libqxcb.so文件即可。原创 2023-07-07 14:55:56 · 7307 阅读 · 11 评论 -
Dynamic dimensions required for input: input, but no shapes were provided. Automatically overriding
pth->onnx->trt原创 2022-06-29 10:26:57 · 881 阅读 · 4 评论 -
The engine plan file is generated on an incompatible device, expecting compute 8.6 got compute 7.5,
出错原因:在2080显卡上将onnx转换成trt模型,2080的算力是7.5然后调用运行trt模型是在3080显卡,算力是8.6,然后就会报错:[E] [TRT] INVALID_CONFIG: The engine plan file is generated on an incompatible device, expecting compute 8.6 got compute 7.5, please rebuild.详细错误信息:解决办法:正如提示,重新在3080上编译,.原创 2022-03-24 11:59:31 · 5023 阅读 · 4 评论 -
‘localtime‘: This function or variable may be unsafe.
vs2019 或者 vs2015解决办法。(1) vs2019_CRT_SECURE_NO_WARNINGS(2)vs2015_CRT_SECURE_NO_WARNINGS原创 2022-03-03 09:41:08 · 437 阅读 · 0 评论 -
D:\ProgramData\Anaconda3\envs\test_onnx\python.exe: No module named pip
window下cmd:python -m ensurepip原创 2022-02-28 17:20:10 · 579 阅读 · 0 评论 -
Could not load library cudnn_cnn_infer64_8.dll.Error code 126 Please make sure cudnn_cnn_infer64_8.
问题场景:c++部署onnx模型当前环境:onnxruntime1.6.0, cuda10.2, cudnn8.3.2原因:cudnn版本不对,改为cudnn8.0.3即可具体环境配置可参考:https://blog.csdn.net/jizhidexiaoming/article/details/122878801https://blog.csdn.net/jizhidexiaoming/article/details/122878801...原创 2022-02-16 15:12:38 · 1351 阅读 · 0 评论 -
ImportError: cannot import name ‘get_all_providers‘ from ‘onnxruntime.capi._pybind_state‘
debug备份场景:pth转onnx遇到此问题环境:onnx 1.8.0onnxruntime-gpu 1.6.0cuda 11.1问题:cuda版本不对尝试解决办法:conda install cudatoolkit=10.2.89成功解决!原创 2022-02-15 15:29:01 · 1293 阅读 · 0 评论 -
Script file ‘D:\Program Files\anaconda3\envs\yudaoai\Scripts\pip-script.py‘ is not present.
报错信息:Script file 'D:\Program Files\anaconda3\envs\yudaoai\Scripts\pip-script.py' is not present.步骤如下:1.下载piphttps://pypi.org/project/pip/#files2.解压到当前文件夹3.进入到此文件夹下cd D:\software\pip-21.3.1D:4.进入到出问题的环境conda activate data_processing5.安装pytho.原创 2021-10-26 22:28:00 · 599 阅读 · 1 评论 -
ERROR: Cannot unpack file C:\Users\ADMINI~1\AppData\Local\Temp\pip-unpack-t6nl10mn\simple
指定源时要加上 "-i" 而不是"i"(win_onnx) C:\Users\Administrator>pip install onnxruntime==1.2.0 i https://pypi.douban.com/simpleCollecting https://pypi.douban.com/simple Using cached https://pypi.douban.com/simple (17.4 MB) ERROR: Cannot unpack file C:\User原创 2021-05-21 11:19:14 · 3704 阅读 · 0 评论 -
undefined reference to `Py_SetPythonHome‘
1.报错信息:undefined reference to `Py_SetPythonHome'undefined reference to `Py_BuildValue'undefined reference to `Py_SetPythonHome'undefined reference to `Py_Initialize'undefined reference to `PyImport_ImportModule'undefined reference to `PyDict_Ge原创 2021-05-12 18:10:42 · 1350 阅读 · 1 评论 -
onnx-tensorrt/builtin_op_importers.cpp:766:12: error: ‘class nvinfer1::IDeconvolutionLayer’ has no m
编译onnxruntime源码时,报错:/home/zxq/cxx/onnxruntime/cmake/external/onnx-tensorrt/builtin_op_importers.cpp:766:12: error: ‘class nvinfer1::IDeconvolutionLayer’ has no member named ‘setDilationNd’ layer->setDilationNd(dilations); ^~~~~~~~~~~原创 2021-05-03 21:20:50 · 1240 阅读 · 0 评论 -
Failed to initialize NVML: Driver/library version mismatch
现象重启电脑后,分辨率不对,使用nvidia-smi命令报错:Failed to initialize NVML: Driver/library version mismatch原因本地自动更新了显卡驱动,重新安装即可。操作# 卸载现有的显卡驱动sudo apt-get purge nvidia*# 添加源sudo add-apt-repository ppa:graphics-driverssudo apt-get update# 找到你适合的版本ubuntu-dri原创 2021-04-25 10:51:57 · 563 阅读 · 0 评论 -
QObject::moveToThread: Current thread (0x5621b4823b40)
完整错误信息:QObject::moveToThread: Current thread (0x5621b4823b40) is not the object's thread (0x5621b6f72df0).Cannot move to target thread (0x5621b4823b40)qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "/home/zxq/anaconda3/envs/open-mmlab/l原创 2021-04-08 13:47:42 · 1472 阅读 · 2 评论 -
unable to get repr for <class ‘torch.Tensor‘>
问题描述:计算BCE Loss使用pytorch接口self.bce_loss = nn.BCELoss()self.bce_loss(pred_cls, tcls)问题原因:pred_cls没有归一化的0到1之间。解决办法:self.bce_loss(torch.sigmoid(pred_cls), tcls)原创 2020-11-02 09:19:45 · 2595 阅读 · 0 评论 -
ModuleNotFoundError: No module named ‘mmcv._ext‘
更改mmcv安装方式:打开:https://github.com/open-mmlab/mmcv点击下图作者已经建立好的对应版本,进行安装。我的是torch1.6. cuda10.2:pip install mmcv-full==latest+torch1.6.0+cu102 -f https://download.openmmlab.com/mmcv/dist/index.htmlInstalling collected packages: mmcv-fullSucce..原创 2020-09-27 22:06:28 · 30351 阅读 · 23 评论 -
磁盘空间为0, 排查大文件
https://www.cnblogs.com/yingsong/p/9468260.html查看磁盘空间#df -mFilesystem 1M-blocks Used Available Use% Mounted on/dev/xvda2 17647 17589 0 100% /tmpfs 1941 0 1941 0% /de...转载 2020-08-21 10:34:25 · 582 阅读 · 0 评论 -
UserWarning: Matplotlib is currently using agg, which is a non-GUI backend, so cannot show the figur
程序的开头加上下面这句:import matplotlibmatplotlib.use('TkAgg')原创 2020-08-03 17:43:33 · 398 阅读 · 0 评论 -
AttributeError: ‘COCO‘ object has no attribute ‘get_cat_ids‘
原因:mmdetection用到了pycocotools,本来接口是pycocotools.coco.getCatIds,但是mmdetection又在外包了一层,变成了get_cat_ids,导致pycocotools.coco又没有这个成员函数,导致出错。解决办法:mmdetection官方建议将coco.py回退到以前版本git checkout 206107 -- mmdet/datasets/coco.py...原创 2020-07-07 14:35:46 · 2352 阅读 · 6 评论 -
python json.load报错:UnicodeDecodeError utf-8 codec
# json_file = open(new_json_file_path, 'r', encoding='GBK') # 这个就不会报错json_file = open(new_json_file_path, 'r', encoding='utf-8') # 报错json.load(json_file)原创 2020-05-19 17:27:25 · 6201 阅读 · 1 评论 -
运行pytorch版faster-rcnn遇到的问题
1,ModuleNotFoundError: No module named 'pycocotools._mask'参考http://www.manongjc.com/article/44682.html摘录如下:1)下载https://github.com/philferriere/cocoapi文件,虚拟环境下cd 到PythonAPI目录下,执行pythonsetup.py...原创 2020-02-21 16:13:35 · 992 阅读 · 0 评论 -
Windows 安装caffe 不能加载libcaffe
最后我的配置的win10, x64, vs 2013, caffe, cuda 7.5, cudnn v5.1,matlab2014b 安装成功。1. 按照网上教程安装好几天,都出现不能加载libcaffe问题,最后发现是cuda版本问题,https://github.com/Microsoft/caffe,这里了下载的CommonSettings.props内配置的cuda版本为7....原创 2018-04-30 15:29:11 · 13972 阅读 · 1 评论 -
配置Pycharm,解决终端可运行,Pycharm报错:Process finished with exit code 134 (interrupted by signal 6: SIGABRT)
错误提示如下:I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcublas.so.8.0 locallyI tensorflow/stream_executor/dso_loader.cc:126] Couldn't open CUDA library libcudnn....原创 2018-07-04 21:50:59 · 34490 阅读 · 9 评论 -
RuntimeWarning: invalid value encountered in arccos
出错代码位置:np.arccos(cos_value) * (180 / np.pi) def get_cross_angle(l1, l2): arr_a = np.array([(l1.p2.x - l1.p1.x), (l1.p2.y - l1.p1.y)]) # 向量a arr_b = np.array([(l2.p2.x - l2....原创 2019-11-13 11:49:52 · 7991 阅读 · 2 评论 -
ECDSA host key for 192.168.16.6 has changed and you have requested strict checking. Host key verific
根据上面提示,解决办法:ssh-keygen -f "/home/pi/.ssh/known_hosts" -R 192.168.16.6ssh zhangxq@192.168.16.6原创 2019-09-26 17:48:16 · 992 阅读 · 1 评论 -
显卡驱动与pytorch版本不匹配
conda安装pytorch后,报显卡驱动版本太低错误。解决办法,更新显卡驱动。ubuntu-drivers devices:sudo add-apt-repository ppa:graphics-drivers/ppasudo apt updatesudo apt install nvidia-430...原创 2019-09-25 15:21:59 · 3609 阅读 · 3 评论