timerring

Stop Slop English writing skill

Tue, 15 Sep 2026 01:37:52 +0800

Stop Slop^[1] is a lightweight, open-source prompt skill for editing English prose that exhibits predictable LLM writing patterns. It provides Markdown instructions and examples for making generated or revised text more direct, concrete, and editorially restrained.

human-writing：让中文写作有点“活人感”

Mon, 14 Sep 2026 22:22:54 +0800

让 AI 写中文，有时候会得到一篇很工整的文章。句子通顺，段落也齐全，就是读着有点累，像在看一份说明书。最近看到一个专门处理这类问题的 Skill，叫 human-writing^[1]，中文名是“活人感写作”。里面主要是 Markdown 写作规则，还有一个 Python 检查脚本。

How to Subscribe to Gemini Pro Outside the US

Thu, 29 Jan 2026 10:03:20 +0800

If you’re located outside the United States and trying to subscribe to Gemini Pro, you might have encountered the frustrating “Not available in your region” message. This guide explains why this happens and how to fix it.

Vim Cheatsheet

Tue, 13 Jan 2026 17:38:50 +0800

Vim is a powerful command-line text editor that’s been around for decades, and it’s still widely used today. Whether you’re SSH-ing into a remote server or just prefer the keyboard-centric workflow, Vim is often the go-to choice. The learning curve can be steep at first — I remember the first time I opened Vim and couldn’t even figure out how to quit. But once you get the hang of it, the efficiency gains are real. This cheatsheet covers the essentials I use most often, and I’m trying to condense them.

ERC 8004 Trustless Agents

Thu, 16 Oct 2025 02:23:05 +0800

This article is my understanding of the ERC-8004: Trustless Agents, which is a standard for Identity Registry, Reputation Registry and Verification Registry. And refer to the Ethereum Improvement Proposals ERC-8004: Trustless Agents^[1].

Fix the Ghost Notification of Github

Mon, 13 Oct 2025 19:06:52 +0800

I ran into a stubborn GitHub notification blue dot that wouldn’t go away even though the Notifications page showed nothing. Here’s the quick, safe way I found to resolve it, plus a few alternatives and notes on why it happens.

Disable SIP on Macos

Sun, 12 Oct 2025 19:39:00 +0800

System Integrity Protection (SIP) is a security layer that limits even the root user from modifying protected parts of the system. Apple introduced it in OS X El Capitan to reduce the blast radius of malware and accidental system damage. In practice, SIP blocks writes to system locations, prevents runtime injection into protected processes, and restricts some kernel/system-level operations.

Sometimes I run a command and see ‘Operation not permitted’ despite sudo, SIP is often the reason. So I need to disable SIP temporarily.

How to Partially Clone

Thu, 21 Aug 2025 21:56:10 +0800

Sometimes you encounter a large repository but only need a few directories from it. Cloning the entire repo wastes both time and disk space. Git’s sparse checkout solves this problem by letting you pull only the parts you need.

The Tips of Macos Automator

Sun, 22 Jun 2025 14:27:16 +0800

The Automator is a powerful tool for automating tasks on macOS. It allows you to create custom workflows that can be triggered by various events, such as opening a file, pasting text, or receiving an email. This article will share my common usages of Automator.

Mainland Network Circuits

Thu, 19 Jun 2025 22:24:36 +0800

The performance and reliability of internet connectivity to and from mainland China are determined by a complex web of different network circuits, each with its own characteristics.

We will explore the crucial differences between the standard public network backbones, like the workhorse 163 network, and premium, optimized lines like CN2. We’ll also explain what dedicated networks such as IPLC are and why they offer a completely different level of performance and privacy. By the end of this article, you’ll have a clear framework for understanding these circuits, helping you make more informed decisions whether you’re choosing a hosting provider, diagnosing network issues, or are simply curious about how the internet infrastructure in mainland China operates.

Cloudflare Warp

Fri, 06 Jun 2025 14:52:52 +0800

Frankly speaking, the first time I heard about Cloudflare is the service so called Cloudflare WARP which is provided by Cloudflare. Nowadays, you must have heard the fastest and private network can be provided by Cloudflare. If I say Cloudflare achieves the fastest object by the first step to visiting any website, the world’s fastest public DNS 1.1.1.1. You can probably guess the private provider, yes, that is Cloudflare WARP. So this article I will introduce the Cloudflare WARP further zero trust and how to use it.

The principle of Cloudflare SaaS Optimization

Wed, 04 Jun 2025 11:23:44 +0800

In common situation, when you use the cloudflare to resolve your domain name, and you choose to use the cloudflare proxy(the yellow cloud icon), then cloudflare would probably get your domain resources into their CDN, that could be a good thing, but as we know, the network in mainland has some special rules, so for the free tier user, the access to the resources in the CDN would be somehow worse, even rest. That means, if the user connect to the node which has high latency, the request will experience worse. So this article will introduce some methods to help your resources access to the cloudflare nodes which are optimized for the mainland.

The MitM and HTTPS Capturing

Tue, 03 Jun 2025 16:03:16 +0800

Do you really know what is the network activity on your phone? Have you ever wondered how to capture the https packages on your phone? In this article, I will introduce you to the MitM and how to capture the https traffic on your phone.

How to Keep Your MAC Always Awake

Fri, 30 May 2025 22:01:25 +0800

Here are some common scenarios: You’re about to step away from your Mac for a period, or you’re downloading large files and want to keep your Mac from going to sleep. (If you use an external display, you likely understand what I mean. Sometimes, after waking from sleep mode, your Mac fails to power on the external display.) That’s why I highly recommend Amphetamine to you. I’ve often heard that macOS boasts excellent power management, but lacking in-depth knowledge myself, I’ve sought out the answer.

The Way to Get Cookie

Sun, 25 May 2025 13:34:21 +0800

本文介绍获取 Cookie 最直接常用的方法。

How to Choose the Right Software and Creative License

Sat, 24 May 2025 20:06:58 +0800

Normally, there are many types of software licenses, each with different terms and conditions. Choosing the right license for your project can be a daunting task. In this article, we will discuss the different types of software licenses and how to choose the right one for your project. And in the end, I make a simple website https://license.timerring.com^[1] to help you choose exactly the license you need.

Add Live Photo Effect to Hugo

Fri, 16 May 2025 16:15:56 +0800

Live Photos 作为最早 Apple 在2015年9月发布的 iPhone6s 和 6s Plus 中引入的特色功能，很快就在手机摄影领域独树一帜，收获了无数忠实用户，Live Photo 的特色在于，它不仅仅是一张静态的照片，用户在拍摄实况照片时，iPhone会录下拍照前后各 1.5 秒所发生的一切。在 24 年下半年，随着微信宣布支持实况照片，Live Photo 再次成为了各大应用以及手机厂商的焦点，在短时间内无法实现更好的 AI 叙事的当下，不少手机厂商将 Live photo 适配各大社交媒体视为这个季度的业绩救赎，最近我也在思考，是否可以在博客中添加这个 feature，以便可以更好地展示照片的内容，在经过一番仔细搜索之后，发现网上完全没有关于 Hugo 支持 Live photo 的任何讨论，因此我决定自己动手实现这个 feature，也就有了本篇文章。

The Practice of Resolving Domain

Sun, 11 May 2025 11:10:46 +0800

According to the last article about A Brief Introduction to DNS, you may have a basic understanding of DNS. In this article, I will show you how to resolve domain name to IP address or other information in practice.

Hidden Markov Models

Mon, 21 Apr 2025 14:49:15 +0800

This is a lecture note of Hidden Markov Models.

Markov Chains

Sun, 20 Apr 2025 18:38:25 +0800

This is the notes for the lecture of Markov Chains.

Markov Processes

Sun, 20 Apr 2025 15:35:34 +0800

This is a summary of the lecture notes of Markov Processes.

Find the Formula Rendering Issue of Hugo

Fri, 18 Apr 2025 13:47:15 +0800

最近在写数学相关的博客的时候，发现 mathjax 的渲染效果不尽人意，经常会在渲染的时候出现非预期的渲染效果，虽然只是极个别的公式，对整体影响并不大，当然我可以放着不解决，继续进行接下来的事情，或者简单将公式通过截图的形式放在文章里，也无伤大雅。但是凡事得想长远，想本质，今天大可以为了这一个公式采用省事的方法解决，那么它不会是我写的最后一个公式，未来可能还会有更多，因为这是一个长期的问题，如果放任不管，可能下次遇到类似的问题，我可能还是会采用省事的方法解决，那么这个问题就会一直存在，所以我认为放下手头的事，集中解决这个问题是值得的。

Multi Platform Builds Docker

Wed, 09 Apr 2025 16:39:42 +0800

This is a viable build record. The multi platform builds are supported by docker buildx^[1], which is related to binfmt_misc and QEMU.

The Invitation Only Mechanism: From Clubhouse to Manus

Fri, 04 Apr 2025 11:40:39 +0800

虽然自小就听到过会员制的概念，但是也仅限于超市或者药店作为一种鼓励积分的方式，亦或是理发店等固定销售半径的捆绑销售的手段。不同于现在人们所谈论“邀请会员制”的概念，关于“邀请会员制”，我认为英文中有一个非常有意思的词，叫做 exclusive，意思是“只，仅仅”，但是在礼物等场景中 exclusive 尝尝用来形容“特定的，独家的”，追根溯源，本质上都是一种排他性(exclusivity)的文化覆盖。

How to Use Git Submodule

Fri, 28 Mar 2025 15:44:52 +0800

Git submodule^[1] is a feature of Git that allows you to include a Git repository as a subdirectory of another Git repository. It is often used to manage dependencies between projects.

Live Streaming infra and protocols

Mon, 24 Mar 2025 23:02:26 +0800

In our daily life, you may always see the live streaming on the internet, such as the live streaming of the video conference, the live streaming of the sports event, the live streaming of the online course, etc. And sometimes you even start your own live streaming, but we all get used to just clicking the button on the screen to start the live streaming, do you know what is happening behind the scene?

Implement danmaku rendering algorithm from scratch

Sat, 22 Mar 2025 13:13:16 +0800

This article presents a comprehensive implementation of a danmaku rendering algorithm from the ground up, along with a thorough analysis of the danmaku rendering algorithm. The source code is available on GitHub^[1].

Is There an Rss Renaissance in the AI Era?

Fri, 21 Mar 2025 13:34:19 +0800

RSS: RDF Site Summary or Really Simple Syndication。

如果说 search 是信息 pull 的代表，那毫无疑问，RSS 就是信息 push 的代表。简单地讲，就是一种信息聚合的服务，通过一种标准化的消息来源的格式规范，聚合多个网站的内容更新并且通知订阅者。尽管听起来很复杂，但是 RSS 本质上只需要两步，从目标网站上找到 RSS 链接，并把它添加到自己内容收取的服务里。

Stock Prices

Thu, 20 Mar 2025 15:16:35 +0800

This is a tutorial for stock prices in the context of random processes.

Possion Processes

Wed, 19 Mar 2025 22:01:16 +0800

This article is mainly about the Poisson processes in the book “Random Processes” by Gregory F. Lawler.

Docker With GPU

Tue, 18 Mar 2025 12:21:21 +0800

This post is to introduce how to use gpu in docker, which also covers the basic usage of gpu on the host machine.

Stochastic Processes

Sat, 15 Mar 2025 19:30:50 +0800

Compared to deterministic model, which is specified by a set of equations that describe exactly how the system will evolve over time. In a stochastic model, the evolution is at least partially random and if the process is run several times, it will not give identical results.

Some Useful Tools

Thu, 13 Mar 2025 15:13:51 +0800

Some useful tools for developers.

Note: This post will be updated regularly.

A Shopping Experience in Pang Donglai Supermarket

Fri, 07 Mar 2025 14:09:23 +0800

前天在广州找朋友时发现附近有一家永辉超市，它是去年年末自主改调的“胖东来”超市，我一下子来了兴致，之前虽然也去过河南很多城市，但是没有去过许昌和新乡(New York)，而胖东来又不在河南之外开店，所以一直缺少体验的机会。在胖东来仔细转了一遍后，我对胖东来的服务有了更进一步的理解和感受。

Langchain and RAG Best Practices

Wed, 05 Mar 2025 00:05:24 +0800

This is a quick-start essay for LangChain and RAG which mainly refers to the Langchain chat with your data^[1] course which are taught by Harrison Chase and Andrew Ng.

Random Walks

Tue, 18 Feb 2025 22:14:11 +0800

This article is a summary of the random walks in the book “Random Processes” by Gregory F. Lawler.

Sequences of Random Variables

Sun, 02 Feb 2025 20:56:05 +0800

Parameter Estimation

Sun, 02 Feb 2025 14:55:34 +0800

This is the notes for the lecture of Parameter Estimation.

Two Random Variables

Sat, 01 Feb 2025 23:55:24 +0800

This article is about the two random variables and their properties.

Further Understanding of Proc

Sat, 01 Feb 2025 20:49:58 +0800

The /proc in Linux is a kind of Pseudo file system (also known as Virtual File System), what it saves is a series of special files that store the current running state of the kernel, and the user can check the hardware info and the runnning processes via the files, even adjust some files to change the state of running kernel. The files in /proc will be refresh in RAM, so the files’ size is 0 bytes.

The Iftop

Thu, 30 Jan 2025 23:57:06 +0800

I frequently use ifconfig and cat /proc/net/dev to examine network information. However, today, my cloud server crashed unexpectedly. I reached out to the cloud support for assistance. The engineer utilized iftop to analyze the situation. After the issue was resolved, I started to consider using this tool. Now, let’s take a closer look at this powerful utility.

Some Useful Commands to Share

Thu, 30 Jan 2025 23:38:01 +0800

Some good commands to share.

Note: This post will be updated regularly.

Random Variables

Tue, 28 Jan 2025 22:15:30 +0800

This article introduces the common distribution functions both continuous and discrete. Besides, the basic knowledge of conditional distribution such as total probability and Bayes’ Theorem is also introduced. The most important part are the Normal Approximation and Poisson Approximation of the Binomial Random Variable.

Repeated Trials

Tue, 28 Jan 2025 15:08:44 +0800

This article is mainly about Bernoulli Trials and its properties.

Probability and Its Axioms

Mon, 27 Jan 2025 19:54:20 +0800

This article is about the probability and its axioms.

Reflections on Trending Topics

Sat, 25 Jan 2025 15:42:49 +0800

一次制造热点帖子引发的思考与总结。

CPU can only see the threads

Thu, 23 Jan 2025 23:36:22 +0800

In python, due to the GIL (Global Interpreter Lock), which is a mutex and ensures only one thread can execute at a time, so the multiple threads parallel execution is not supported under the CPython interpreter. But what about the multiple processes? What is the difference between them? How to choose the right method? Do you know the coroutine? Let’s explore it together.

Go Common Test

Wed, 22 Jan 2025 17:08:47 +0800

The test methods in Go mainly include three types: unit test, benchmark test, and example test.

Go Concurrency and Parallelism

Wed, 22 Jan 2025 14:00:45 +0800

Go Cheatsheet

Tue, 21 Jan 2025 18:30:07 +0800

Video Technology 101

Mon, 20 Jan 2025 11:01:09 +0800

Do you really understand the video technology? What is the frame rate, resolution, bit rate and their relationships? Do you know the parameters p and K in the terminology 1080p and 4K? Do you know how to describe the video quality? What is the Blu-ray Disc on earth? What the difference between the codec such as H.264/AVC, H.265/HEVC, AV1? You may notice the Apple ProRes in apple, so do you really understand what it for? Do you know the HDR and Dolby Vision? So why there are so many tailers such as .mp4, .mkv, .flv, etc?

Now this Video Technology 101 will introduce them to you.

The Main Kind of Message Queue

Sun, 19 Jan 2025 20:15:13 +0800

Message queues are a form of asynchronous service-to-service communication. They are important in enhancing a system’s scalability, reliability, and maintainability.

The Method to Manage Traffic

Sun, 19 Jan 2025 17:33:27 +0800

Do you know the basic method to manage the traffic? There are four methods: Load Balancers, Reverse Proxies, Forward Proxies, and API Gateways. And they have different features and usage scenarios.

The Different Kind of API Design

Sun, 19 Jan 2025 14:42:18 +0800

In the daily life, we often hear about different kind of API, such as REST API, GraphQL API, WebSocket, Webhook, RPC and gRPC even the SOAP, so do you know the difference between them? How to choose the right API for your project?

The Encode and Decode in Python

Sun, 19 Jan 2025 13:07:49 +0800

Do you really know the encode and decode in Python?

About the Systemd

Sat, 18 Jan 2025 22:06:14 +0800

It has been a long time since the linux used init to manage the startup process, such as sudo /etc/init.d/apache2 start or service apache2 start, but the init is serial. To address this issue, the systemd was born. The d is the abbreviation of daemon, which means the systemd is a daemon manager. The systemd substitutes the initd and becomes the default main process PID 1.

The Tips About Dockerfile

Sat, 18 Jan 2025 21:05:23 +0800

Normally, we often write a Dockerfile in the current directory.

The Dockerfile is a configuration file that describes how to build the image. You can refer to the official documentation^[1] for more details.
If you list more than one CMD, only the last one takes effect. So if you have multiple commands to run, you better write them in a script file.
Docker is not the VMware, there is no systemd in the container. Its startup program is the container application process. The container exists for the main process. Once the main process exits, the container loses its meaning of existence and thus exits. So when you execute multiple commands and if they are blocking, you better write the previous commands in nohup and the last command in the blocking command. (never use the command such as CMD service nginx start, the CMD only will execute as CMD [ "sh", "-c", "service nginx start"], when the sh is executed, the container will exit, the correct way is run it directly CMD ["nginx", "-g", "daemon off;"])

Docker Cheatsheet

Fri, 17 Jan 2025 10:06:35 +0800

This is a cheatsheet of docker.

Note: the docs will be updated from time to time.

Docker 101

Thu, 16 Jan 2025 22:33:49 +0800

Docker is a practical tool for everyday use, and like Git, you can learn it in just 30 minutes.

The Instance Class Static Magic Method in Python

Tue, 14 Jan 2025 21:51:24 +0800

So what is the difference between the Instance method, the Class method and the Static method?

The Review and Plan for Bilive

Fri, 10 Jan 2025 19:24:37 +0800

从 12 月 10 号正式上线生产版本的 bilive 到现在，已经过去了一个月了。目前 bilive 的 star 数已经突破了 300，实际用户数应该在 20 ～ 30 之间。当然 star 的数据本身没有什么意义，我真正追求的是使用者，但是 star 数据能够反映出项目实际浏览量，与其他工具类项目相比，我对于 bilive 的定位更像是一个产品，如何让这个产品积累足够多的用户和反馈，是我目前主要面临的问题。

The Overview of Security

Wed, 08 Jan 2025 20:42:43 +0800

In today’s digital age, security is paramount. As we increasingly rely on technology for communication, commerce, and data storage, understanding the fundamentals of security becomes essential. This article provides an overview of key security concepts, including encryption, digest algorithms, and digital signatures. By exploring these topics, we aim to equip you with the knowledge to protect your digital assets and ensure the integrity and confidentiality of your information. Whether you’re a tech enthusiast or a professional in the field, this guide will offer valuable insights into the mechanisms that safeguard our digital world.

How to Publish Your Code as a Pip Module

Tue, 07 Jan 2025 20:46:08 +0800

Last week, I have made a python cli tool. To make it more convenient to use, I want to publish it as a pip module, so I have made some research and mistakes, and finally succeeded.

Git Cheatsheet

Tue, 07 Jan 2025 20:39:35 +0800

This is a Git cheatsheet that distills my years of experience.

Workspace: The directory of the repository. The working directory is independent of each branch.
Staging Area (Index): A temporary area for data, similar to a cache before writing from the working directory to the repository. The staging area is independent of each branch.
Repository: Stores all code versions that have been committed to the local repository.
Version Structure: A tree structure where each node represents a code version.

Some Good Things to Share About the Packages

Mon, 06 Jan 2025 19:59:20 +0800

As a developer, I have been using many packages in my projects. Sometimes I may even forget the existence of the packages. When I changed my devices or some one else asked me, it’s hard for me to remember everything to recommend. Thus I will try to list the things which I think are very helpful and useful to share with you. Maybe we can even make an “Annual Oscar Awards ceremony” for the things.

Note: This post will be updated regularly.

Introduction to the HTTP and HTTPS Protocol

Sat, 04 Jan 2025 15:15:31 +0800

HTTP protocol is the foundation of the Internet, and HTTPS is the secure version of HTTP. HTTP is an application layer protocol based on TCP/IP protocol. It does not involve packet (packet) transmission, mainly specifying the communication format between the client and the server, and the default port is 80.

Mail Service and Protocol

Sat, 04 Jan 2025 10:53:02 +0800

Recently, I have been working on DNS of my domain name. And then I need to set up the mail service of my domain name. When I tried many times, I always failed to receive the mail on my phone. Suddenly, I remembered that I didn’t set up IMAP service. So let’s review the mail service and protocol.

How to Add Commits to Someone Else's Pull Request

Fri, 03 Jan 2025 17:06:19 +0800

Most of time, we need to reply the commiter to adjust some code, sometimes it’s not efficient. Instead, we can directly add the commit to the PR.

How to Merge Multiple Commit to the One

Fri, 03 Jan 2025 17:05:15 +0800

Sometimes we make too many small commits during development, and it’s better to squash them into one clean commit before pushing. Here’s how to use git rebase -i to merge multiple commits.

Understanding Clash Through Configuration

Thu, 02 Jan 2025 15:11:55 +0800

We can learn the proxy process of clash through the configuration file.

Thinking About Advertisement from an Open Source Perspective

Fri, 27 Dec 2024 15:17:09 +0800

我在本科期间听到过很多关于广告论调，仍记得某位教授对于短视频所代表的互联网深恶痛绝，在课堂上批判广告的价值，抨击广告行业的浮躁。我身边的同学也大抵持相同的观点，认为互联网广告就是智商税，广告的目的是为了卖课，卖课的目的是为了收割。可事实真的如此吗？

A Brief Introduction to DNS

Thu, 26 Dec 2024 21:10:14 +0800

In my mind, DNS is the key of the internet. I always believe if you control the DNS, you control the Internet world. So let us get started to know the DNS.

Real Computer Network

Mon, 23 Dec 2024 20:06:18 +0800

For the computer network, I believe that most people have learned it in the course of university or college. No matter whether you are familiar with it or not, this article will give you a different perspective of the network. It can really help you a lot.

Python Generator Iterator and Decorator

Mon, 23 Dec 2024 19:19:14 +0800

This article will review the basic knowledge of Python, including generator, iterator, and decorator.

Python Underlying Mechanism

Mon, 23 Dec 2024 19:12:31 +0800

This article will review the underlying mechanism of Python, including data types and derivation operations.

Python Files Exceptions and Modules

Mon, 23 Dec 2024 19:02:35 +0800

This article will review the basic knowledge of Python, including files, exceptions, and modules.

Python Object Oriented Programming

Mon, 23 Dec 2024 18:23:06 +0800

This article will review the object-oriented programming of Python. Mainly about the class, inheritance, and polymorphism.

Python Cheatsheet

Sun, 22 Dec 2024 00:09:26 +0800

Python Function Parameter

Sat, 21 Dec 2024 22:09:09 +0800

This blog will review the function parameter of Python. Mainly about the parameter passing, Keyword and Position Parameter, Anonymous Function, and unit test with assert.

Python Control Structure

Sat, 21 Dec 2024 20:26:15 +0800

This blog will review the control structure of Python.

Python Composite Data Type

Sat, 21 Dec 2024 19:38:25 +0800

This article will review the composite data type in Python, such as list, tuple, dictionary, and set.

Python Basic Data Type

Sat, 21 Dec 2024 16:51:57 +0800

This article will explore data types and common methods for them in Python. The content of this review is as follows: number type, string type, boolean type and type conversion.

Python Basic Syntax Elements

Fri, 20 Dec 2024 23:19:25 +0800

Sometimes you may forget about the basic of Python, so let us take a look at the summary of Python basic summary. The content of this review is as follows: data type, variable, control flow, input & output, PEP8 format.

Deploy Github Pages With GPG Signing

Wed, 18 Dec 2024 23:25:21 +0800

I have been busy migrating my blog this week. Coincidentally, I learned that there may be cases of commit forgery on GitHub. Therefore, for security reasons, I added a GPG signature. However, when deploying Hugo, I encountered many problems regarding whether GPG signatures can also be used. Fortunately, I finally solved them.

GPG 101

Tue, 17 Dec 2024 20:26:57 +0800

GnuPG is a complete and free implementation of the OpenPGP standard as defined by RFC4880(also known as PGP). GnuPG allows you to encrypt and sign your data and communications; it features a versatile key management system, along with access modules for all kinds of public key directories. GnuPG, also known as GPG, is a command line tool with features for easy integration with other applications.

This article I will introduce the basic usage of GPG.

Housewarming 2024

Sat, 14 Dec 2024 14:53:36 +0800

So after a long time, I decide to restart my blog program. In my daily development, I have a lot of thoughts and ideas, thus I will write some documents to record them. But I believe that sharing is the best way to learn. A specific example is cryptography, which means the closing source algorithm is never the safest. Only algorithms that have been vetted by the public are truly secure.

The macOS Cheatsheet

Wed, 28 Aug 2024 20:00:00 +0800

整理常用的 macOS 键盘快捷键与触控板操作。这里的 ⌘ 表示 Command，⌥ 表示 Option，⌃ 表示 Control；有些组合键的效果取决于当前应用及系统设置。

Building a RAG Chatbot with NVIDIA AI Endpoints

Sun, 18 Aug 2024 10:15:14 +0800

基于 NVIDIA AI Endpoint、LangChain 和 FAISS 构建 RAG 智能对话机器人，实践商品问答与检索增强生成。

Python Linear Algebra

Sun, 14 Jul 2024 20:00:00 +0800

Scalar #

Scalar variables are denoted by ordinary lowercase letters (for example, $x$, $y$, and $z$). $\mathbb{R}$ denotes the space of all (continuous) real-valued scalars, and $x\in\mathbb{R}$ is the formal way to say that $x$ is a real-valued scalar. A scalar is represented by a tensor with only one element.

Python Data Preprocessing

Fri, 12 Jul 2024 23:17:46 +0800

Reading the Dataset #

The data is stored in the CSV (comma-separated values) file ../data/house_tiny.csv.

import os

os.makedirs(os.path.join('..', 'data'), exist_ok=True)
data_file = os.path.join('..', 'data', 'house_tiny.csv')
with open(data_file, 'w') as f:
 f.write('NumRooms,Alley,Price\n') # Column names
 f.write('NA,Pave,127500\n') # Each row represents one data sample
 f.write('2,NA,106000\n')
 f.write('4,NA,178100\n')
 f.write('NA,NA,140000\n')

# To load the raw dataset from the CSV file we created, import `pandas` and call `read_csv`.
# !pip install pandas
import pandas as pd

data = pd.read_csv(data_file)
print(data)

# NumRooms Alley Price
# 0 NaN Pave 127500
# 1 2.0 NaN 106000
# 2 4.0 NaN 178100
# 3 NaN NaN 140000

Python Data Manipulation

Fri, 12 Jul 2024 22:58:44 +0800

An $n$-dimensional array is also called a tensor. Regardless of the deep learning framework, its tensor class (ndarray in MXNet and Tensor in PyTorch and TensorFlow) is similar to NumPy’s ndarray. During deep learning training, we usually do not read images one at a time. Instead, we might read 128 images at once. This is a batch, represented by a four-dimensional array. Deep learning frameworks also offer important features beyond NumPy’s ndarray:

Strong support for GPU-accelerated computation, while NumPy supports only CPU computation;
Automatic differentiation in tensor classes.

A tensor with one axis corresponds to a mathematical vector; A tensor with two axes corresponds to a mathematical matrix.

My Must Haves for Macos Memo

Sat, 22 Jun 2024 15:53:14 +0800

This article will share my must-haves for Macos in order to make my Macos more efficient and comfortable.

Data Export

Sun, 25 Feb 2024 19:47:10 +0800

1. Export data #

Since R is mainly used for data analysis, import files are more commonly used than export files, but sometimes we also need to export data or analysis results. The functions write.table( ) and write.csv( ) can export data to a .txt file and a .csv file, respectively.

In addition, the function save( ) can save the specified object in the workspace as an R data file with the extension .rdata. For example:

ID <- 1:5
sex <- c("male", "female", "male", "female", "male")
age <- c(25, 34, 38, 28, 52)
pain <- c(1, 3, 2, 2, 3) 
pain.f <- factor(pain, levels = 1:3, labels = c("mild", "medium", "severe")) 
patients <- data.frame(ID, sex, age, pain.f)
save(patients, file = "patients.rdata")

# IMPORT DATA
load("patients.rdata")

the rdata format file takes up less space and loads quickly with R. Therefore, it is recommended that users ** save data in rdata format ** after importing and organizing data in other formats. To import data in this format, simply call the load( ) function.

Strategies for Processing Large Datasets

Sun, 25 Feb 2024 19:47:08 +0800

In real-world problems, data analysts may face datasets with hundreds of thousands of records and hundreds of variables. Processing such large datasets requires a relatively large amount of computer memory, so use a 64-bit operating system and a device with relatively large memory whenever possible. Otherwise, data analysis may take too long or may even be impossible to carry out. In addition, effective strategies for processing data can greatly improve analysis efficiency.

11 Multithreading and Locks

Sat, 13 Jan 2024 19:57:43 +0800

11.1 Multithreading #

11.1.1 Implementing Multithreading #

Approach 1: Extend the Thread class.

class Worker extends Thread {
 @Override
 public void run() {
 for (int i = 0; i < 10; i ++ ) {
 System.out.println("Hello! " + this.getName());
 try {
 Thread.sleep(1000);
 } catch (InterruptedException e) {
 throw new RuntimeException(e);
 }
 }
 }
}

public class Main {
 public static void main(String[] args) {
 Worker worker1 = new Worker();
 Worker worker2 = new Worker();
 worker1.setName("thread-1");
 worker2.setName("thread-2");
 worker1.start();
 worker2.start();
 }
}

10 Annotations and Reflection

Fri, 12 Jan 2024 21:32:06 +0800

10.1 Annotations #

(1) Annotations, also called metadata, describe packages, methods, fields, constructors, local variables, and other program elements.
(2) Annotations do not affect program logic, but can be processed at compile time or runtime.
(3) In Java SE, annotations serve relatively simple purposes, such as marking obsolete features or suppressing warnings. In Java EE, they play a more important role, such as configuring aspects of an application in place of the verbose code and XML configuration left over from older Java EE versions.

09 Exception Handling

Thu, 11 Jan 2024 11:37:58 +0800

1. Differences Between Error and Exception #

An Error is a failure that a program cannot handle, such as OutOfMemoryError or ThreadDeath. When these occur, the Java Virtual Machine (JVM) generally terminates the thread. Such failures are fatal to the program and cannot be caught and handled.

An Exception is an exceptional condition that the program itself can handle. Exceptions fall into two categories: runtime and non-runtime exceptions. Programs should handle these exceptions whenever possible.

08 Common Containers

Wed, 10 Jan 2024 20:49:57 +0800

8.1 List #

Interface: java.util.List<>.

Implementations:

java.util.ArrayList<>: resizable array
java.util.LinkedList<>: doubly linked list

Methods:

add(): add an element at the end
clear(): clear the list
size(): return its length
isEmpty(): check whether it is empty
get(i): get the ith element
set(i, val): set the ith element to val

Loop syntax:

// Standard
for(int i = 0; i < list.size(); i ++) { 
 System.out.println(list.get(i)); 
}
// Enhanced
for (Integer integer : list) { 
 System.out.println(integer); 
}

07 Classes and Interfaces

Tue, 09 Jan 2024 22:24:11 +0800

7.1 Classes and Objects #

A class defines an entirely new data type containing a group of variables and functions; an object is an instance of that class.

Defining Student as a class represents the abstract concept of a “student.” Each student is then an object (instance) of the Student class.

06 Functions

Mon, 08 Jan 2024 19:18:32 +0800

Function Basics #

A typical function definition includes the following parts: modifiers, return type, function name, a list of zero or more parameters, and the function body.

05 Strings

Sun, 07 Jan 2024 21:53:46 +0800

1. Characters and Integers — ASCII Codes #

Each commonly used character corresponds to a number in -128 ~ 127, and the two can be converted to each other. Note: currently, negative numbers have no corresponding characters.

import java.util.Arrays;

public class Main {
 public static void main(String[] args) {
 char c = 'a';
 System.out.println((int)c);

 int a = 66;
 System.out.println((char)a);
 }
}

Common ASCII values: 'A'- 'Z' are 65 ~ 90, 'a' - 'z' are 97 - 122, and 0 - 9 are 48 - 57.

In Java, character literals must be enclosed in single quotes and string literals in double quotes. They are distinct. (Python and JS do not make this distinction.)

Characters can participate in arithmetic; they are treated as integers during operations:

import java.util.Arrays;

public class Main {
 public static void main(String[] args) {
 int a = 'B' - 'A';
 int b = 'A' * 'B';
 char c = 'A' + 2;

 System.out.println(a); // 1
 System.out.println(b);
 System.out.println(c);
 }
}

04 Arrays

Sat, 06 Jan 2024 20:36:19 +0800

One-Dimensional Arrays #

Defining Arrays #

Array definitions are similar to variable definitions.

public class Main {
 public static void main(String[] args) {
	 int[] a;// Definition: define an array a
		a = new int[10]; // Initialization: initialize a as an array of length 10, meaning a contains 10 int variables; equivalent to =>int[] a = new int[10]// Define and initialize at the same time
 int[] a = new int[10], b;
 float[] f = new float[33];
 double[] d = new double[123];
 char[] c = new char[21];
 // Arrays can be defined for more than just primitive data types; String arrays can also be defined
 String[] strs = new String[10];// This array contains 10 String variables
 }
}

03 Loops

Fri, 05 Jan 2024 22:07:53 +0800

Use == to compare primitive data types for equality; for other types (such as strings and functions), use equals to compare them.

02 Conditional Statements

Thu, 04 Jan 2024 19:42:08 +0800

If Statements #

if-else statements can also contain other if-else statements.

import java.util.Scanner;

public class Main {
 public static void main(String[] args) {
 Scanner sc = new Scanner(System.in);
 int a = sc.nextInt(), b = sc.nextInt(), c = sc.nextInt();

 if (a > b) {
 if (a > c)
 System.out.println(a);
 else
 System.out.println(c);
 } else {
 if (b > c)
 System.out.println(b);
 else
 System.out.println(c);
 }
 }
}

Chained if-else statements

Enter a score between 0 and 100.
If it is at least 85, output A;
if it is at least 70 but less than 85, output B;
if it is at least 60 but less than 70, output C;
if it is less than 60, output D.

import java.util.Scanner;

public class Main {
 public static void main(String[] args) {
 Scanner sc = new Scanner(System.in);
 int s = sc.nextInt();

 if (s >= 85) {
 System.out.println("A");
 } else if (s >= 70) {
 System.out.println("B");
 } else if (s >= 60) {
 System.out.println("C");
 } else {
 System.out.println("D");
 }
 }
}

01 Variables, Operators, Expressions, Input and Output

Wed, 03 Jan 2024 21:14:37 +0800

Syntax Basics #

Variables #

In Java, outputting a variable without assigning it a value causes an error.
Built-in data types

Type	Bytes	Example	Additional notes
byte	1	123
short	2	12345	$2^{16}=65536$
int	4	123456789	$2^{32}=2147483648≈2.1*10^9$
long	8	1234567891011L	$2^{64}≈10^{18}$
float	4	1.2F	About 6 or 7 significant decimal digits; subsequent digits are imprecise approximations. To explicitly denote a float, append f or F.
double	8	1.2, 1.2D	About 15 or 16 significant decimal digits; subsequent digits are imprecise approximations. To explicitly denote a double, append d or D.
boolean	1	true, false
char	2	‘A’	ASCII code correspondences: a-97 A-65 0-48

Density Plotting

Thu, 31 Aug 2023 22:19:04 +0800

Seaborn’s kdeplot() function is one way to draw density plots in Python. At present, Matplotlib does not have a specific function for drawing density plots, and plotting is generally performed in combination with the result of the gaussian_kde() function in the Scipy library.

The neighbors.KernelDensity () module in Python’s scikit-learn library provides six kernel functions—Gaussian, Tophat, Epanechnikov, Exponential, Linear, and Cosine—for kernel density estimation calculations. Python’s KDEpy library provides as many as nine kernel functions, including Gaussian, Exponential, Box, Tri, Epa, Biweight, Triweight, Tricube, Cosine.

A schematic diagram of the nine kernel functions in the KDEpy library is shown below:

Types of Univariate Plots for Continuous Variables

Tue, 29 Aug 2023 18:32:03 +0800

A univariate plot (chart for one variable) refers to drawing a corresponding plot using one variable from a dataset. To visualize this variable, a plot must be drawn according to the different data variable types. Data variables are divided into continuous variables and discrete variables.

Message Queues and Kafka Basics

Sun, 27 Aug 2023 21:57:58 +0800

走进消息队列 #

面对存储行为服务崩溃的时候，可以采用解耦的方式，将记录加入消息队列中，然后作为生产者，此时存储服务作为一个消费者来消耗。

面对多个请求同时进行时，可以采用削峰的方式解决，将订单也加入消息队列中，然后每次只处理 10 个请求。

把发起订单的请求放入消息队列中，这时就可以通知客户下单成功了，然后将下面的三个请求异步地执行即可。

将日志放入消息队列中，然后通过日志的相关组件例如 ES 引擎等等，Kibana 展示平台进行分析。

TOS Object Storage: Fundamentals and Architecture

Sun, 27 Aug 2023 17:20:40 +0800

抖音背后的存储 #

短视频生产/消费：片源系统->审核系统->推荐系统
公共系统：客户端、账号系统、评论系统

处理流程：

手机上传视频到片源服务，片源服务将视频存储。
片源服务将视频发送抽帧服务和转码服务。
抽帧服务将视频抽帧采样并存储。
转码服务将原视频转换为不同清晰度、不同码率的视频并存储。
抽帧后的图片被送往审核服务。
机器或人工审核过后视频被传给推荐服务。
用户可以获取推荐列表或下载推荐视频/图片。

存储量分析：

因此存储要求：易用、海量、便宜。

SciencePlots Basics and Features

Sat, 26 Aug 2023 16:38:14 +0800

Users sometimes need to customize layer properties such as fonts, tick axes, spines, and legends according to a journal’s figure-drawing requirements, which is time-consuming and can also easily cause users to overlook some layer-detail requirements.

As a third-party extension toolkit specifically for drawing scientific-paper figures, SciencePlots provides Matplotlib plot styles (Matplotlib Styles) for mainstream English-language scientific journals (such as Nature, Science, and IEEE).

The installation code for SciencePlots is as follows:

pip install SciencePlots

ProPlot Basics and Features

Fri, 25 Aug 2023 09:46:57 +0800

Figures for scientific papers impose higher requirements on the rendering conditions of multilayer elements (fonts, axes, legends, etc.), and we need to change multiple rendering parameters in Matplotlib and Seaborn, especially when drawing complex figures containing multiple subplots, which can easily make the plotting code verbose.

As a concise Matplotlib wrapper, the ProPlot library is a high-level wrapper for Matplotlib’s object-oriented plotting method (object-oriented interface), integrating the cartopy/Basemap mapping libraries, xarray, and pandas, and can compensate for some of Matplotlib’s shortcomings. ProPlot can give Matplotlib enthusiasts a smoother plotting experience.

Seaborn Basics and Features

Thu, 24 Aug 2023 23:37:29 +0800

Seaborn provides a more advanced encapsulation on top of Matplotlib, enabling users to use very little code to draw a visualization library for scientific paper figures with rich statistical information. Seaborn is based on Matplotlib, most parameters of plotting functions in Matplotlib can be used in Seaborn plotting functions, and it provides good support for other Python libraries (such as Numpy/Pandas/Scipy).

Install Matplotlib

python -m pip install matplotlib

Install Seaborn

pip install seaborn

Foundations of Scientific Figure Design and Color

Wed, 23 Aug 2023 17:12:42 +0800

Fundamentals of Drawing Scientific Paper Figures #

Drawing Specifications #

1.Classification and Composition of Scientific Paper Figures #

Example of the basic composition of a scientific paper figure

Plotting Tools and Their Key Features

Tue, 22 Aug 2023 18:41:14 +0800

Python has multiple third-party libraries, including the basic plotting library Matplotlib, the statistical analysis visualization library Seaborn, the scientific plotting library ProPlot, and the scientific plotting theme library SciencePlots.

Numerical Descriptive Analysis

Sun, 20 Aug 2023 22:56:22 +0800

Before the analysis, first convert the categorical variables low, race, smoke, ht, and ui in the birthwt dataset into factors.

library(MASS)
data(birthwt)
str(birthwt)

options(warn=-1)
library(dplyr)
birthwt <- birthwt %>% 
 mutate(low = factor(low, labels = c("no", "yes")),
 race = factor(race, labels = c("white", "black", "other")),
 smoke = factor(smoke, labels = c("no", "yes")),
 ht = factor(ht, labels = c("no", "yes")),
 ui = factor(ui, labels = c("no", "yes")))
str(birthwt)

ggplot2 and Other Plots

Sat, 19 Aug 2023 22:19:09 +0800

1. Getting to Know the ggplot2 Package #

The ggplot2 package provides a plotting system based on a grammar of layers. It remedies the lack of consistency among functions in R’s base graphics system and raises R’s plotting capabilities to an entirely new level. The basic principles of the various data visualizations in ggplot2 are completely consistent: it maps mathematical space to the space of graphical elements. Imagine a blank canvas. On the canvas, we need to define the data to be visualized (data) and the mapping from data variables to graphical properties (mapping).

Plotting with R's Base Graphics System

Sat, 19 Aug 2023 21:33:26 +0800

R’s base graphics system was written by Ross Ihaka and is very powerful. It consists mainly of the graphics and grDevices packages, which are loaded automatically when R starts. There are two types of functions in the base graphics system: high-level plotting functions and low-level plotting functions.

High-level plotting functions are functions used to produce plots directly, including plot( ), hist( ), boxplot( ), and pairs( ). Low-level plotting functions are functions used to add new graphics or elements to a plot drawn by a high-level plotting function, including points( ), lines( ), text( ), title( ), legend( ), and axis( ).

Handling Missing Values

Sat, 19 Aug 2023 20:03:10 +0800

Handling Missing Values #

In actual data analysis, missing data are frequently encountered. Missing values usually occur because data were not collected or were not entered.

For example, a missing age may be due to someone not providing his (her) age. Most statistical analysis methods assume that complete datasets are being processed. Therefore, apart from some specialized books, most statistics textbooks rarely address this issue. In fact, before conducting a formal analysis, we need to check whether the dataset contains missing values during the data preparation stage and use some methods to compensate for the loss caused by missing values.

Merging Data Frames

Sat, 19 Aug 2023 14:19:21 +0800

Sometimes datasets come from multiple places, and we need to merge two or more datasets into one dataset. Operations for merging data frames include vertical merging, horizontal merging, and merging by a shared variable.

Using the dplyr Package

Fri, 18 Aug 2023 21:35:27 +0800

This package handles data frames more efficiently with a unified specification. ** The first parameter of all functions that process data frames in the dplyr package is the data frame name. **

Taking the birthwt dataset in the MASS package as an example, the following describes the use of commonly used functions in the dplyr package. This data set comes from a case-control study of risk factors for low birth weight in newborns. Start by loading the dataset and viewing its related information.

library(dplyr)
data(birthwt, package = "MASS")
# ??birthwt

A total of 189 subjects and 10 variables were included in the dataset birthwt. where the outcome variable bwt is the weight of the newborn (unit: g) and the variable low is a binary classification variable that converts the value of bwt into 2500g points. The remaining 8 variables are predictors, including pregnant women’s age, race, smoking status, and history of hypertension (ht).

Using Base R

Thu, 17 Aug 2023 23:05:23 +0800

Using Base R #

In actual data analysis, analysts often spend a lot of effort on data preparation, transforming data into the form required for analysis. Unfortunately, most statistics textbooks rarely address this important issue. Organizing data is one of the tasks of statistics. We started to focus on the most common data format in R - the basic operation of data frames. We will first process the data frames using the base package.

Load a small dataset Familydata in the epiDisplay package first.

library(epiDisplay)
data("Familydata")

Data Acquisition

Thu, 17 Aug 2023 21:27:32 +0800

In fact, R has a large number of built-in datasets available for analysis and practice, and we can also create data in R that simulate specific distributions. In actual work, however, data analysts more often face external data from various data sources, namely data files with all kinds of extensions, such as .txt, .csv, .xlsx, .xls, and so on. Files with different extensions represent different file formats, which often causes trouble for analysts.

R provides data import tools with a wide range of applications.

Organizing Code with Design Patterns

Thu, 17 Aug 2023 20:00:00 +0800

好的设计模式方便对代码进行功能扩展/修改，降低耦合性，提高复用性。

from abc import abstractmethod, ABC
from typing import Callable, final, Tuple, List, Sequence, Dict
from random import choices, randrange
from functools import wraps
from numpy import ndarray
from pandas import DataFrame
from time import perf_counter, sleep
from datetime import datetime, timedelta

Data Structures and Conversion

Wed, 16 Aug 2023 20:08:02 +0800

The first step in any data analysis is to create a dataset in the required format. In R, this task consists of two steps: first select a data structure to store the data, and then enter or import the data into this data structure. The following introduces the various data structures used to store data in R.

Dividing Functionality with Classes

Wed, 16 Aug 2023 17:15:09 +0800

理解类是运用设计模式的前提；运用类本身也能更好的组织代码，提升扩展性、可读性、复用性。

from typing import Callable, Tuple, List, Sequence
from pandas import DataFrame

Writing Maintainable Code

Wed, 16 Aug 2023 16:00:00 +0800

from typing import Callable, Tuple, List
from pandas import DataFrame

项目结构定好了，就开始着手写代码了；那编写代码有哪些注意点呢？

Introduction to R

Tue, 15 Aug 2023 22:19:41 +0800

Common Arithmetic Operators #

Operator	Description
+	Addition
−	Subtraction
*	Multiplication
/	Division
^ or **	Exponentiation
%%	Remainder
%/%	Integer division

Note that R is case-sensitive.

Building a Project Structure

Tue, 15 Aug 2023 20:00:00 +0800

在具体的数据科学场景下，按照项目结构→代码编写→功能划分→组织代码的顺序，组织高质量的代码。

项目结构也可认为是目录结构，例如

Image Name

我们将项目代码拆分成多个功能模块，固化成函数/类并写到不同的文件中，存放在不同的路径里，python 将每个 py 文件视作为一个模块，可以用 import 调用；例如在 main.py 里调用 /db/db.py 里的函数 create_conn

from db.db import create_conn 
conn = create_conn() # create_conn创建了一个数据库的链接对象

CUDA Matrix Multiplication and Transpose

Thu, 10 Aug 2023 11:59:16 +0800

通过代码实践 CUDA 矩阵乘法与转置，理解 GPU 存储单元与并行计算。

CUDA Thread Indexing

Wed, 09 Aug 2023 10:19:58 +0800

结合向量相加和图像处理，介绍 CPU 与 GPU 内存操作、线程索引以及 Grid 和 Block 的配置。

我们要处理的问题

如图，有离散型的，例如向量相加等等。有 local 卷积类型的，用其他一片的元素值求得一个元素。还有 All to All 类型的，类似于傅里叶变换。

回顾一下 CUDA 程序的编写：

把输入数据从 CPU 内存复制到 GPU 显存
在执行芯片上缓存数据，加载 GPU 程序并执行
将计算结果从 GPU 显存中复制到 CPU 内存中

Apache Thrift RPC and Matchmaking Service

Tue, 08 Aug 2023 21:11:31 +0800

thrift 是一种接口描述语言和二进制通讯协议，它被用来定义和创建跨语言的服务。

Thrift 是 Facebook 于 2007 年开发的跨语言的 RPC（Remote Procedure Call，远程过程调用）服务框架，提供多语言的编译功能，并提供多种服务器工作模式；用户通过 Thrift 的 IDL（接口定义语言）来描述接口函数及数据类型，然后通过 Thrift 的编译环境生成各种语言类型的接口文件，用户可以根据自己的需要采用不同的语言开发客户端代码和服务器端代码,让不同语言构建的服务可以做到远程调用无缝对接。（它们的目的都是提供本地调用远程服务的能力）

RPC（Remote Procedure Call，远程过程调用）是一个计算机通信协议，此协议允许进程间通信。简单来说，当机器 A 上的进程调用机器 B 上的进程时，A 上的调用进程被挂起，而 B 上的被调用进程开始执行。调用方可以通过参数将信息传送给被调用方，然后可以通过被调用方传回的结果得到返回
大部分的RPC框架都遵循如下三个开发步骤：

定义一个接口说明文件：描述了对象(结构体)、对象成员、接口方法等一系列信息；

通过RPC框架所提供的编译器，将接口说明文件编译成具体的语言文件；

在客户端和服务器端分别引入RPC编译器所生成的文件，即可像调用本地方法一样调用服务端代码；
RPC是一种C/S架构的服务模型，server端提供接口供client调用，client端向server端发送数据，server端接收client端的数据进行相关计算并将结果返回给client端。

Go Practical Examples: Guessing Game, CLI Dictionary, and SOCKS5 Proxy

Wed, 02 Aug 2023 21:36:33 +0800

这篇文章完整整理自本地的 Go 语言实战入门笔记，依次练习猜数字、命令行词典和 SOCKS5 代理。保留原笔记的分步讲解、代码及外部图床图片。

运行前注意：词典案例使用第三方接口，当前可用性及授权方式尚未验证。原笔记的浏览器令牌和 Cookie 已从代码中移除；运行前应从服务提供方取得合法凭据，并在本机环境变量 CAIYUN_DICT_TOKEN 中设置，切勿将凭据提交到仓库。SOCKS5 示例没有认证，仅用于本机学习，请勿作为公网代理部署。图片链接指向原笔记的外部图床，是否可长期访问取决于图床。

LaTex basic concepts

Wed, 02 Aug 2023 16:50:13 +0800

TEX #

TEX is software developed by Donald E. Knuth for typesetting text and mathematical formulas. In 1977, while writing The Art of Computer Programming, Knuth realized that the steadily declining typesetting quality would affect the publication of his work. To reverse this situation, he began developing TEX, tapping into the potential of digital printing equipment that had just begun to be used in the publishing industry at the time. In 1982, Knuth released the TEX typesetting engine, and later improved it in 1989 to better support 8-bit characters and multilingual typesetting. TEX is renowned for its outstanding stability, cross-platform capabilities, and almost bug-free nature. Its version number continuously approaches π, and is currently 3.141592653.

Writing Your First CUDA Program

Wed, 02 Aug 2023 11:53:41 +0800

从第一个 CUDA 程序开始，实践 NVCC 编译、Makefile、多文件编译、线程索引和性能分析。

CUDA Programming Model

Fri, 28 Jul 2023 20:27:45 +0800

梳理 CUDA 异构计算、执行空间说明符、线程层次、NVCC 编译流程和 NVProf 性能分析。

GPU Hardware Architecture

Fri, 28 Jul 2023 18:13:08 +0800

介绍 CPU 与 GPU 的架构差异、GPU 硬件平台、Jetson 设备及 Amdahl 定律。

Introduction to CUDA and Linux Basics

Mon, 24 Jul 2023 21:09:38 +0800

从 CUDA 与 NVIDIA 技术生态入门，介绍 Ubuntu、Jetson 开发环境、权限管理、SSH 和 Makefile。

AI开发硬件环境综述

Tue, 14 Mar 2023 17:55:58 +0800

笔记本选配 #

出于通勤等因素，建议型号

MacBook Air M1 16+ 256
MacBook Pro M1 16 + 256

M1芯片的mbp非常强大，发热不严重，甚至在Air版没风扇.

Overview of the AI Development Software Environment

Tue, 14 Mar 2023 16:47:12 +0800

Regarding software:

Installation of the ubuntu operating system
Abandon bash, embrace zsh
Use of package managers,
Install the NVIDIA GPU driver
Software installation:Miniconda
Software installation:Python IDE
Software installation: ssh
Software installation:Git
Installation of other productivity software

Nginx Basics and Minimal Configuration

Wed, 01 Mar 2023 20:35:29 +0800

Directory Structure #

The Nginx root directory contains the following folders:

client_body_temp conf fastcgi_temp html logs proxy_temp sbin scgi_temp uwsgi_temp

Folders ending in _temp are used to store temporary files generated during operation.

The other main folders are:

conf: stores configuration-related files
html: the default directory for static files such as HTML and CSS files
sbin: the main Nginx executable
logs: stores various logs. For example, access records access-related information, error records errors, and nginx.pid records the service’s PID, that is, its process ID.

Nginx Installation and Deployment

Tue, 28 Feb 2023 21:57:27 +0800

Virtual machine: VMware Workstation 16 Operating system: CentOS 7.4 Download URL: https://vault.centos.org/centos/7.4.1708/isos/x86_64/CentOS-7-x86_64-Minimal-1708.iso^[1]

Recommended computer configuration:

Memory: 8 GB or more is recommended
Disk: An SSD is recommended
CPU: Any mainstream CPU with four or more cores will suffice

SQL Query Optimizers Explained

Fri, 05 Aug 2022 20:38:20 +0800

从 SQL 解析与执行流程出发，介绍基于规则和代价的查询优化、统计信息、执行计划枚举，以及 Apache Calcite 的开源实践。

Pipes, Environment Variables, and Common Commands

Wed, 27 Jul 2022 20:00:00 +0800

管道类似于文件重定向，可以将前一个命令的stdout重定向到下一个命令的stdin。

管道命令仅处理 stdout，会忽略 stderr （标准错误）。
管道右边的命令必须能接受stdin。
多个管道命令可以串联。

Tmux Common Commands

Tue, 26 Jul 2022 20:00:00 +0800

A tmux instance can contain multiple sessions, a session can contain multiple windows, a window can contain multiple panes, and a shell dialog is opened in each pane.

Shell Scripting Fundamentals

Mon, 25 Jul 2022 20:00:00 +0800

Shell is the language we use to communicate with the operating system through the command line.

Shell scripts can run directly on the command line, or a set of logic can be organized into a file for reuse. The command line can be thought of as a “shell script executing line by line.”

There are many common kinds of shell scripts in Linux, including:

Bourne Shell (/usr/bin/sh or /bin/sh)
Bourne Again Shell (/bin/bash)
C Shell (/usr/bin/csh)
K Shell (/usr/bin/ksh)
zsh
…

Linux systems generally use bash by default.

The file should begin with #! /bin/bash to specify bash as the script interpreter.

Introduction to CUDA

Fri, 08 Jul 2022 11:58:01 +0800

介绍 CUDA 的基本概念、支持的编程语言、发展历程及 GPU 加速计算的应用。

MySQL Cheatsheet

Tue, 15 Mar 2022 03:43:29 +0800

整理 MySQL 常用语法、函数、约束、多表查询和事务，并介绍存储引擎、索引、执行计划与 SQL 性能优化。文中的密码仅作示例，实际使用时请独立生成。

Image Classification and Foundational Vision Models

Mon, 07 Mar 2022 16:42:54 +0800

Process #

Collect data
Define the model

Usually a function containing parameter variables: $y=F_{\Theta}(X)$

Example: $y=\sigma\left(\Theta^{T} X\right)$
Train Find the optimal parameters $\Theta^{*}$ so that the model $y=F_{\Theta^{*}}(X)$ achieves the highest accuracy on the training set
Predict For a new image $\hat{X}$ , use the trained model to predict its class, namely $\hat{y}=F_{\Theta^{*}}(\hat{X})$

SSH Remote Access and File Transfer

Sat, 12 Feb 2022 20:10:20 +0800

Log In to the Server #

ssh user@hostname

user: username
hostname: IP address or domain name

The first time you log in, you will see this prompt:

The authenticity of host '123.57.47.211 (123.57.47.211)' can't be established.
ECDSA key fingerprint is SHA256:iy237yysfCe013/l+
Are you sure you want to continue connecting (yes/no/[fingerprint])?

Enter yes and press Enter. This records the server’s information in ~/.ssh/known_hosts.

Then enter the password to log in to the remote server. The default login port is 22. To log in through a specific port:

ssh user@hostname -p 22

Introduction to Web APIs

Sat, 10 Jul 2021 21:28:31 +0800

Strictly speaking, most of what we learned in the JavaScript section belongs to ECMAScript. ECMAScript, or ES, defines language standards such as the syntax rules for variables, data types, expressions, statements, and functions. Browsers implement most of the ECMAScript specification and extend it with practical features. These extensions are called Web APIs.

JavaScript Data Types

Fri, 09 Jul 2021 23:01:26 +0800

JavaScript is a weakly typed, or dynamic, language. You do not need to declare a variable’s type in advance; its type is determined automatically while the program runs.

var age = 10; // Number
var areYouOk = 'yes'; // String

JavaScript has dynamic types, meaning the same variable can be used for different types (its type can change).

var x = 6; // x is a number
var x = "Bill"; // x is a string

JavaScript Objects

Fri, 09 Jul 2021 20:00:00 +0800

Objects are a JavaScript data type. We have already learned about numbers, strings, booleans, and undefined. An object can be understood as a collection of data. It consists of properties and methods.

JavaScript Functions

Fri, 09 Jul 2021 16:58:16 +0800

Declaration (Definition) #

Declaring (defining) a complete function involves five parts: the keyword, function name, formal parameters, function body, and return value.

JavaScript Arrays

Fri, 09 Jul 2021 16:00:00 +0800

Arrays: An array is a data type that stores data in order.

JavaScript Operators and Control Flow

Fri, 09 Jul 2021 15:07:15 +0800

Commonly used operators in JavaScript include:

Arithmetic operators
Increment and decrement operators
Comparison operators
Logical operators
Assignment operators

JavaScript Fundamentals

Wed, 07 Jul 2021 16:30:35 +0800

Brendan Eich (born 1961) completed the design of JavaScript in 10 days.

It was originally named LiveScript and was later renamed JavaScript after a collaboration with Sun.
JavaScript is one of the world’s most popular languages. It is a scripting language that runs on the client side (“script” means a script), and now it can also be used for server-side programming with Node. Js technology.
Scripting language: it does not need to be compiled; during execution, the js interpreter (js engine) interprets and executes it line by line.

HTML Basics

Mon, 05 Jul 2021 20:00:00 +0800

A web page consists of three parts:

Structure: written using HTML
Presentation: CSS is used to set the styles of elements on the page
Behavior: JavaScript is used to define the page’s behavior HTML is HyperText Markup Language. It is responsible for the page’s structure and defines its various parts. HTML is written as plain text and uses HTML tags to identify the different parts of a page.

About

Mon, 01 Jan 0001 00:00:00 +0000

About #

John Howe
GPG Public: 0x26FEE0805E6C9F71
Key fingerprint = B613 FBAF 0822 BEBD FABF 8F65 26FE E080 5E6C 9F71

If you want to know what this blog is about, here are some representative posts:

Network Real Computer Network | A Brief Introduction to DNS ｜ Introduction to the HTTP and HTTPS Protocol
Develop CPU can only see the threads | Python series
Audio and Video Video Technology 101 | Implement danmaku rendering algorithm from scratch
Security The Overview of Security | GPG 101
Container Docker 101 | Docker Cheatsheet

Coffee chat #

blog topic and coffee chat buy me a coffee

Bookmarks

Mon, 01 Jan 0001 00:00:00 +0000

Bookmarks

No need to copy someone else’s collection. The best things are the ones that stay with you.

Stop Slop

Sep 12, 2026 Skills

Human Writing

Sep 12, 2026 Skills

OpenAI Developers

Sep 9, 2026 Product

Google Research

Sep 9, 2026 Research

Device Archive

Mon, 01 Jan 0001 00:00:00 +0000

Device Archive

2026

NOW

In Use

Band

WHOOP Life

2026 — Present

WHOOP MGECGScreen-Free

“Remarkably accurate automatic activity recognition and recovery tracking.”

In Use

Eyeglasses

Friends

Mon, 01 Jan 0001 00:00:00 +0000

Followed blogs #

Google Research Cloudflare OpenAI OpenAI Developers Claude Cursor Windsurf Perplexity Sourcegraph Cerebras Mixedbread Morphllm tavily

Friends links #

Nodeseek Sukka Selfboot zu1k Spencer Woo zhile.io 52txr Colah Bojie Li Bimu Timothy immmmm Harry Chen Qcrao Liam ishell dewx Eryajf Darmau gaoweix limbopro Changqian Yu ice1000 lsvih zhheo jiejoe shuyuej likun only.rs aigonna bulianglin greatdk idoubi Jianlin Su axiaoxin theowenyoung kyth zyzhang naiba lixueduan gatesnotes ezyang yuanji taoshu tangwudi tisonkun vitalik cmliussss Sebastian Raschka Marcin Wichary YaGe