特别的猫

python graphviz可视化

解决分割网络的结果一片黑的问题

cudart64_100.dll;cublas64_100.dll等没有找到的解决办法

ValueError: some of the strides of a given numpy array are negative. This is currently not supported

VOC数据集-修改标签

Pytorch错误：Torch not compiled with CUDA enabled

注意力机制

自适应激活函数的实验结果记录

目标检测中常用的激活函数代码实现-pytorch1.2.0

VOC数据集提取自己需要的类

ImportError: TensorBoard logging requires TensorBoard with Python summary writer installed.

训练出错:OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized.解决方案：import osos.environ["KMP_DUPLICATE_LIB_OK"]="TRUE"

RuntimeError: (PreconditionNotMet) The third-party dynamic library (cublas64_102.dll；cublas64_10.dll

小目标检测的数据增强------Stitcher和Mosaic效果对比

将num_workers 修改为0

比赛网址链接：https://www.heywhale.com/home/competition/605ab78821e3f6003b56a7d8 水下目标检测资源：https://github.com/wangdongdut/Underwater-Object-Detection首先看看原始数据集：1、直接拿原始数据训练，YOLOV4 线上：经过图像增强的视觉效果：采用cascade_rcnn_r50_fpn训练图像增强后的数据集，并删除数据集中水草“waterweeds”这一类.

ValueError: Error when checking input: expected input_1 to have 4 dimensions, but got array with shape (1, 512, 512)解决方案：test_img = cv2.cvtColor(test_img, cv2.COLOR_GRAY2BGR)出现问题原因：彩色图和灰度图之间得通道数问题...

UnknownError (see above for traceback): Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.解决方案： #---------------------------------------------.

路径：E:\Anaconda3\Lib\site-packages\tensorboardrun：tensorboard --logdir=dirpath --host=127.0.0.1

您定义的输入形状是单个样本的形状。而对于单个图像32*32的RGB图像，您应该返回一个形状(1,32,32,3)。也就是说输入应该是4维的，长，宽，通道三个维度外加一个图片数量维度，一张照片也应该是4维。...

问题：RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location='cpu' to map your storages to the CPU.解决方案：model.load_state_

数据集：PubLayNet训练：code：import osimport sysimport randomimport torchimport torchvisionfrom torchvision.models.detection.faster_rcnn import FastRCNNPredictorfrom torchvision.models.detection.mask_rcnn import MaskRCNNPredictorfrom torchvision.

error：Traceback (most recent call last): File "C:/Users/Administrator/Desktop/PubLayNet-maskrcnn/PubLayNet/maskrcnn/infer.py", line 117, in <module> main() File "C:/Users/Administrator/Desktop/PubLayNet-maskrcnn/PubLayNet/maskrcnn/infer.py

pip install labelImg -i https://pypi.tuna.tsinghua.edu.cn/simpleCMD 输入：labelImg

文档版面分析：输出文档图片中图、表、标题、文本的位置和文本信息Document layout analysis: Output the position and text information of the graph, table, title, text in the document picture在传统的文档OCR识别技术中，算法会先分析图片中有几个布局区域，然后分析出水平横向文字，竖向垂直文字，表格和配图照片等区域，然后在针对各自的特点进行切分字符，保留区域类型，进行OCR识别调整；所以可

配置github开源的代码，实现了表格框线的识别，需要更近一步的优化处理。The configuration of github open source code, the realization of table box line recognition, the need for further optimization.

通过USB摄像头拍摄的图像非常非常非常模糊，想通过图像增强的方式变得清晰一点，效果不怎么好，看实验结果：The image taken by the USB camera is very, very fuzzy. I want to make it clearer through image enhancement, but the effect is not very good. See the experimental results:Code：import cv2 as cv#全局.

（1）env：pip install vtk（2）test：import vtk# 1. 读取数据cube = vtk.vtkCubeSource()cube.Update()#记得加这句不加看不到模型# 2. 建图（将点拼接成立方体）cube_mapper = vtk.vtkPolyDataMapper()cube_mapper.SetInputData(cube.GetOutput())# 3. 根据2创建执行单元cube_actor = vtk.vtkActor()cube

实现了双目摄像头的实时视频显示，网上有很多相关的资料，但是不一定是可以用的，我现在使用的这个方法是可以用的。Realize the real-time video display of binocular camera, there are a lot of relevant information on the Internet, but it is not necessarily available, I am using this method can be used.看看效果：result

对于局域网通信无非就是传输视频或者进行信息交互，今天介绍一下局域网视频传输的构建。For LAN communication is nothing more than transmission of video or information exchange, today to introduce the construction of LAN video transmission.首先，看看我的程序目录： First, my program directory: config.ini.

绘制采集温度的温度热力场。参考博客：https://blog.csdn.net/liyazhou0215/article/details/72491993/ Draw the thermal field of temperature for the collection temperature. code：import matplotlib.pyplot as pltimport matplotlib.cm as cmfrom matplotlib.colors import LogNo

接着上一篇博客继续实现PDF中表格的绘制：https://blog.csdn.net/m0_37690102/article/details/107051058Then on a blog to continue to implement drawing in PDF form:https://blog.csdn.net/m0_37690102/article/details/107051058直接上代码：在上一篇博客中，您已经能够简单的能在PDF中生成相应的表格了，但是还是存在一点点小问题，对于表.

今天，使用python生成PDF文档，然后在生成的PDF中创建自定义表格。Today, you generate PDF documents in Python, and then create custom tables in the generated PDF.（1）使用python生成pdf文档需要的最基本的包是pdfgen。它属于reportlab模块，而reportlab模块并没有默认集成到python的安装包中，所以需要安装该模块。（2）pdfgen提供了一个绘画操作的对象canvas

Pytorch在win系统下比较难安装，直接使用pip 安装不能正常的安装，您可以参考我的方式。Pytorch is difficult to install in win system. Direct PIP installation cannot be normally installed. You can refer to my method.pip3 install torch==1.2.0+cpu torchvision==0.4.0+cpu -f https://download.p.

参考这一篇博客：https://blog.csdn.net/m0_37690102/article/details/106975159，今天，接着实现上面一篇博客没有实现的活体检测功能，今天，实现这部分功能，界面设计等功能就不在阐述了，主要实现活体检测的功能，使用dlib的库。Today, reference this blog: https://blog.csdn.net/m0_37690102/article/details/106975159, then to realize the above

使用python打印model模型及pb模型的输入输出节点。参考下面的代码：The Input Output nodes of the Model model and THE PB model are printed in Python.Refer to the following code:from tensorflow.python import pywrap_tensorflowimport oscheckpoint_path=os.path.join('checkpoints/crnn

首先，设计UI界面，如图所示： First, design the UI interface, as shown in the figure 使用语句：pyuic5 OboardCamDisp.ui -o OboardCamDisp.py将UI文件编译为PY文件Use: Pyuic5 oboardcamdisp.uI-o oboardcamdisp.py to compile the UI file into py file接下来就是美化界面，参考下面的代码： The next s

参考博客：https://www.cnblogs.com/jfdwd/p/11425487.htmlimport ：import matplotlib.pyplot as plttest code：plt.figure(figsize=(10, 10))plt.plot(x, y, marker='o')ADD text：txt = ['A','V1','V2','V3','V4','V5','V6','V7','V8','V9','V10','V11','V12']f.

通过Python实现指定函数的时间测试，比较有用的一个代码，您可以直接复制使用。Through The Python implementation of the specified function time test, a more useful code, you can directly copy the use.import time__all__ = ['print_time']def print_time(f): """Decorator of viewing func

虚线处理成直线，这个代码是很有效的。 Dashed lines are processed into straight lines, and this code is very efficient. INPUT：OUTPUT：代码暂时不公开，请见谅 The code is temporarily closed, please forgive me I hope I can help you,If you have any questions, pleasecomm...