king52113141314的博客

原创 Contrastive Learning

https://towardsdatascience.com/understanding-contrastive-learning-d5b19fd96607 The Beginner’s Guide to Contrastive Learning Contrastive Representation Learning | Lil'Log

原创 DCA1000 ROS

在socket编程前，有必要知道大小端模式。大小端模式是指计算机内存中数据存放的模式，由CPU决定，与操作系统无关。大端模式（Big-endian，是指数据的高字节保存在内存的低地址中，数据的低字节保存在内存的高地址中，这种存储模式有点儿类似于把数据当作字符串顺序处理：地址由小向大增加，而数据从高位往低位放；小端模式（Little-endian），是指数据的高字节保存在内存的高地址中，而数据的低字节保存在内存的低地址中，这种存储模式将地址的高低和数据位权有效地结合起来，高地址部分权值高，低地址部分权值低。

原创 Github 搜索技巧

【代码】Github 搜索技巧。

原创 Python GUI图形界面编程

python图形界面开发教程：第一课：编写第一个界面。_哔哩哔哩_bilibiliGitHub - PySimpleGUI/PySimpleGUI: Launched in 2018. It's 2023 and PySimpleGUI is actively developed & supported. Create complex windows simply. Supports tkinter, Qt, WxPython, Remi (in browser). Create GUI applicati

转载 RNN文字情感分析

本章介绍了不同种类的神经网络，有多层神经网络（MLP），卷积神经网络（CNN）和长短记忆模型（LSTM）。它们的共同点是有很多参数，需要通过后向传播来更新参数。CNN 和LSTM 作为神经网络的不同类型的模型，需要的参数相对较少，这也反映了它们的一个共性：参数共享。这和传统的机器学习原理很类似：对参数或者模型加的限制越多，模型的自由度越小，越不容易过度拟合。反过来，模型参数越多，模型越灵活，越容易拟合噪声，从而对预测造成负面影响。

原创 Matlab 简记

1， Find2，

转载 回归（Regression）是什么意思

它并不是存在那里的一个向量，而是描述向量之间的关系。就像我们的大脑一样，我们的大脑中并没有一个像电脑一样的存储介质​​来存储记忆。而是，通过数量众多的神经元，这个世界的时候，我们并不能得到这个世界的全部信息（更加准确，更能精确地描述自变量和因变量之间的。我们的认知又是什么呢？就是回归方程的参数。），只能得到这个世界展现出的可被我们观测的。，就只能通过尽可能多的信息，从而使得我们的。通过更多的数据（自变量和因变量），使得。就是自变量和因变量之间的。那么，如果我们想得到世界的。，并且用回归方程描述。...

原创 Visual Studio Code Keyboard shortcuts for Linux

Visual Studio Code: Multi-line and multi-cursor editing | Ken CenerelliMulti-Line EditingMulti-line editing is similar to column mode editing in Notepad++. The idea is to edit multiple instances of text on contiguous lines and replace it with similar i

原创 自动驾驶中的坐标系

旋转坐标系_zhoucoolqi的博客-CSDN博客_旋转坐标系激光雷达基础：坐标系_zhoucoolqi的博客-CSDN博客_激光雷达坐标系自车坐标系下的物体相对和绝对位置和速度计算_zhoucoolqi的博客-CSDN博客_自车坐标系...

原创 Create bird‘s-eye view using inverse perspective mapping

Create bird's-eye view using inverse perspective mapping - MATLAB3D旋转变换——欧拉角 - 知乎

原创 load your own custom image dataset

https://github.com/aladdinpersson/Machine-Learning-Collection/tree/master/ML/TensorFlow/Basics/tutorial18-customdata-imageshttps://www.analyticsvidhya.com/blog/2021/07/step-by-step-guide-for-image-classification-on-custom-datasets/https://github.com/N.

原创 K-Means & Autoencoder & DBSCAN

Implementing K-Means in Octave/Matlab - Aaqib SaeedGaussian Mixture Models Tutorial and MATLAB Code · Chris McCormickCS221GitHub - powerplayer9/Matlab-KMeans: K-Means Algorithm for RGB Color Imagehttps://github.com/matzewolf/kMeansK-Means Clu...

原创 Evaluating a Classification Model & Training tricks for DL

ROC, AUC, confusion matrix, and metricsEvaluating a Classification Model | Machine Learning, Deep Learning, and Computer VisionUnderstanding perplexity, bits-per-character, and cross entropyEvaluation Metrics for Language Modeling

原创 PointNet++

Furthest Point Sampling | Minibatch AIdef fps(points, n_samples): """ points: [N, 3] array containing the whole point cloud n_samples: samples you want in the sampled point cloud typically << N """ points = np.array(poin...

原创 ROS CV_Bridge编译

Cyan Infinite - Compiling ROS cv_bridge with Python 3Compiling ROS cv_bridge with Python 3Wanted to compile ROS packages with Python 3 on ROS Melodic, only realizing that catkin_make only compiles Python 2 scripts instead after a while? No worries! In

原创 make python script executable in linux

How to create a Linux executable file using python code - Stack Overflowallow it to be executed. To do it, you must use thechmodcommand as follows:chmod +x youfile.py.The+xadd the right to be executed.Ubuntu doesn't know that he must use the...

原创 Kernel Function&Convolution&Correlation&HOG

机器学习有很多关于核函数的说法，核函数的定义和作用是什么？ - 知乎

原创 Image Derivative , optical flow and Computer Vision

When taking the derivative of an image, you’re actually taking what’s called a discrete derivative, and it’s more of an approximation of the derivative. One simple example is that you can take the derivative in the x-direction at pixel x1 by taking the di.

原创 目标检测和跟踪

带你入门多目标跟踪（三）匈牙利算法&KM算法 - 知乎

原创 朴素贝叶斯和SVM

详解最大似然估计（MLE）、最大后验概率估计（MAP），以及贝叶斯公式的理解_nebulaf91的博客-CSDN博客_最大后验估计如何简单理解贝叶斯决策理论（Bayes Decision Theory）？ - 知乎参数估计 | Python 与机器学习...

原创 Vanishing Points and Their Applications

原创 Linux和Windows命令行指令以及环境变量设置

Ashellis a program that takes commands from files or entered interactively through the keyboard and passes them on to the operating system to be executed. A shell is accessed through aterminal (emulator). Where you would normally execute a command by c...

原创 Colab + Vs Code + GitHub + Jupyter (Perfect combination)

https://colab.research.google.com/github/JayThibs/jacques-blog/blob/master/_notebooks/2021-09-27-connect-to-colab-from-local-vscode.ipynb#scrollTo=WA2MmkhS0Prd这个必须看,Caltech课程包含python语法介绍,latex,colab使用,数据处理:BE/Bi 103 a: Introduction to Data Analysis in

转载 pandas和 numpy常用指令总结

pandas常用指令 | citisy的炼丹房1. 数据导入1.1. pd.DataFrame()data，ndarray (structured or homogeneous), Iterable, dict, or DataFrame输入数据源 ndarray：np.array() Iterable：[[1,2,3],[4,5,6]] dict：{'a':[1,2,3], 'b':[4,5,6]}->df.columns=['a','b], df...

原创 Ubuntu使用teamviewer连接windows电脑

1，下载安装teamviewerTeamViewer Linux Download for Remote Desktop access and collaborationsudo apt install -f teamviewer-host_15.25.5_amd64.deb如报错没有权限，命令前加 sudoapt - Unable to locate package -- when the /file/ is right in the directory (it's a .deb) .

原创 accuracy coverage precision recall f1

Evaluation Metrics | Machine Learning | python-course.eu多分类模型Accuracy, Precision, Recall和F1-score的超级无敌深入探讨 - 知乎

原创 When to Use MLP(ANN), CNN, and RNN Neural Networks

Artificial Neural Network (ANN)is also known as aFeed-Forward Neural networkbecause inputs are processed only in the forward direction.ANN vs CNN vs RNN | Types of Neural NetworksWhen to Use MLP, CNN, and RNN Neural NetworksWhen to Use Multilayer...

原创 理解dropout

https://towardsdatascience.com/simplified-math-behind-dropout-in-deep-learning-6d50f3f47275Dropout Regularization in Neural Networks: How it Works and When to Use ItHow to Reduce Overfitting With Dropout Regularization in Kerasanswer is n*m*s...

原创 使用C语言获取Linux系统当前时间

1，精确到毫秒gettimeofday（）//可精确到微秒clock_gettime（）函数，与gettimeofday类似，但精确度到更高，可以精确到纳秒C语言获取Linux系统当前时间（精确到毫秒）_Linuxer_Martin-CSDN博客_linuxc获取系统时间毫秒c语言中获取当前时间的函数_modi000的博客-CSDN博客_c语言获取当前时间的函数2，Convert from epoch to human-readable dateEpoch Converter...

原创 解决shapes (none,111) and (none,111) are incompatible

1,sparse_categorical_crossentropy和categorical_crossentropy的区别对应解决这类问题:“logits and labels must have the same first dimension, got logits shape [32,28] and labels shape [3360]".出现这个错误的原因就是网络最终输出shape和label的shape不一致My features and labels matrices have

原创 Python命令行脚本文件传参和函数传参以及解析参数,spyder,spacy使用

1,Python命令行脚本文件传参运行Python脚本时在命令行中传入参数_guofei_fly的博客-CSDN博客命令行运行Python脚本时传入参数的三种方式 - 汉尼拔草 - 博客园命令行运行Python脚本时传入参数的三种方式_Daniel2333的博客-CSDN博客_python 命令行传入参数2,Python函数传参Python传入参数的几种方法_清风阁-CSDN博客_python 传参...

原创 神经网络的调参方法以及验证集(开发集)的概念

1, 验证集(开发集)的概念训练集(Training set)——用来训练算法的数据集,亦即用来确定W、b参数,训练集就是用来训练参数的，说准确点，一般是用来梯度下降的。 开发集(development set)——用来对学习算法调整参数，选择特征或者做其他的决策的数据集。亦即用来选择超参数（网络层数，网络节点数，迭代次数，学习率这些）.也叫做交叉验证集(hold-out cross validation set) 测试集——既不参与参数的学习过程，也不参与参数的选择过程,仅仅用来评估算法表现(..

原创 TensorFlow2.0学习以及pytest测试指令

简单粗暴 TensorFlow 2 | A Concise Handbook of TensorFlow 2 — 简单粗暴 TensorFlow 2 0.4 beta 文档https://blog.csdn.net/qq_40643699/category_10545446.htmlhttps://blog.csdn.net/qq_38251616/category_10732944.html

原创 Ubuntu 配置Python环境(包括Tensorflow)

1，安装minicondahttps://gist.github.com/arose13/fcc1d2d5ad67503ba9842ea64f6bac35How to Install miniconda on linux (from the command line only)# Setup Ubuntusudo apt update --yessudo apt upgrade --yes# Get Miniconda and make it the main Python inte.

原创 A “single-layer“ perceptron can‘t implement XOR

Single-layer Neural Networks (Perceptrons)

原创 非常好的神经网络直观解释的网站

Neural Networks from Scratch - an interactive guide3Blue1Brown

原创 Ubuntu安装，配置全教程

1，安装开机logo界面狂按 F2，在exit界面选择第三个U盘，然后开机Windows10安装ubuntu18.04双系统教程 - 不妨不妨，来日方长 - 博客园Ubuntu18.04/20.04完整新手安装教程 - 简书https://segmentfault.com/a/11900000221025702，删除firefox改为安装Chrome1）解决dbkg报错：kevalen@kevalen-911:~$ dbkg --get-selections |grep ...

原创 Segmentation fault when running rviz on wsl in Windows10

问题:RViz报错~ rosrun rviz rviz QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-benik' [ INFO] [1571763679.870059200]: rviz version 1.13.4 [ INFO] [1571763679.871324300]: compiled against Qt version 5.9.5 [ INFO] [1571763679.87205..

原创 ROS安装简易教程-ubuntu

melodic/Installation/Ubuntu - ROS Wiki1,Configure your Ubuntu repositoriesConfigure your Ubuntu repositories to allow "restricted," "universe," and "multiverse."Repositories/CommandLine - Community Help Wiki2, sources.list怎么保存和修改sudo nano /etc/.

原创 WSL 从 C 盘迁移至 D 盘的简单实用方法(仅在powershell中操作即可)

1, 安装chocoSet-ExecutionPolicy Bypass -Scope Process -Force; [System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072; iex ((New-Object System.Net.WebClient).DownloadString('https://chocolatey.org/