You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
528 lines
21 KiB
528 lines
21 KiB
<?xml version="1.0" encoding="utf-8"?>
|
|
<feed xmlns="http://www.w3.org/2005/Atom">
|
|
<title>Nicksxs's Blog</title>
|
|
<icon>https://nicksxs.me/icon.png</icon>
|
|
<subtitle>What hurts more, the pain of hard work or the pain of regret?</subtitle>
|
|
<link href="https://nicksxs.me/atom.xml" rel="self"/>
|
|
|
|
<link href="https://nicksxs.me/"/>
|
|
<updated>2025-12-28T13:58:34.706Z</updated>
|
|
<id>https://nicksxs.me/</id>
|
|
|
|
<author>
|
|
<name>Nicksxs</name>
|
|
|
|
</author>
|
|
|
|
<generator uri="https://hexo.io/">Hexo</generator>
|
|
|
|
<entry>
|
|
<title>重新复习理解下java的类加载器</title>
|
|
<link href="https://nicksxs.me/2025/12/28/%E9%87%8D%E6%96%B0%E5%A4%8D%E4%B9%A0%E7%90%86%E8%A7%A3%E4%B8%8Bjava%E7%9A%84%E7%B1%BB%E5%8A%A0%E8%BD%BD%E5%99%A8/"/>
|
|
<id>https://nicksxs.me/2025/12/28/%E9%87%8D%E6%96%B0%E5%A4%8D%E4%B9%A0%E7%90%86%E8%A7%A3%E4%B8%8Bjava%E7%9A%84%E7%B1%BB%E5%8A%A0%E8%BD%BD%E5%99%A8/</id>
|
|
<published>2025-12-28T13:58:34.000Z</published>
|
|
<updated>2025-12-28T13:58:34.706Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前比较粗浅的写过一点类加载器的相关知识，最近因为在看相关的内容，所以打算来复习也简单分享下<br>首先是我们常规用户自定义类的类加载器，</p>
|
|
<figure class="highlight java"><table><tr><td</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="Java" scheme="https://nicksxs.me/categories/Java/"/>
|
|
|
|
|
|
<category term="Java" scheme="https://nicksxs.me/tags/Java/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>记录下ffmpeg命令</title>
|
|
<link href="https://nicksxs.me/2025/12/21/%E8%AE%B0%E5%BD%95%E4%B8%8Bffmpeg%E5%91%BD%E4%BB%A4/"/>
|
|
<id>https://nicksxs.me/2025/12/21/%E8%AE%B0%E5%BD%95%E4%B8%8Bffmpeg%E5%91%BD%E4%BB%A4/</id>
|
|
<published>2025-12-21T14:16:35.000Z</published>
|
|
<updated>2025-12-21T14:16:35.543Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>偶尔会用工具下载网上视频，然后有些下载会是视音频分离的模式，当文件名的开头是”-“的时候，在文件合并的时候就会出现异常<br>因为ffmpeg会将”-“视为命令参数的一部分，所以需要把文件重命名后，手动进行合并<br>命令也比较简单</p>
|
|
<figure</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="工具" scheme="https://nicksxs.me/categories/%E5%B7%A5%E5%85%B7/"/>
|
|
|
|
|
|
<category term="工具" scheme="https://nicksxs.me/tags/%E5%B7%A5%E5%85%B7/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>尝试学习理解下Claude Code的实现原理和细节</title>
|
|
<link href="https://nicksxs.me/2025/12/14/%E5%B0%9D%E8%AF%95%E5%AD%A6%E4%B9%A0%E7%90%86%E8%A7%A3%E4%B8%8BClaude-Code%E7%9A%84%E5%AE%9E%E7%8E%B0%E5%8E%9F%E7%90%86%E5%92%8C%E7%BB%86%E8%8A%82/"/>
|
|
<id>https://nicksxs.me/2025/12/14/%E5%B0%9D%E8%AF%95%E5%AD%A6%E4%B9%A0%E7%90%86%E8%A7%A3%E4%B8%8BClaude-Code%E7%9A%84%E5%AE%9E%E7%8E%B0%E5%8E%9F%E7%90%86%E5%92%8C%E7%BB%86%E8%8A%82/</id>
|
|
<published>2025-12-14T14:33:08.000Z</published>
|
|
<updated>2025-12-14T20:34:19.090Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>结合Claude的回答我们来看下Claude Code的实现原理和细节</p>
|
|
<h3 id="一、诞生故事-从「听歌小工具」到企业级产品"><a href="#一、诞生故事-从「听歌小工具」到企业级产品" class="headerlink"</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>在Antigravity中对比体验Claude Sonnet 4.5模型</title>
|
|
<link href="https://nicksxs.me/2025/12/07/%E5%9C%A8Antigravity%E4%B8%AD%E5%AF%B9%E6%AF%94%E4%BD%93%E9%AA%8CClaude-Sonnet-4-5%E6%A8%A1%E5%9E%8B/"/>
|
|
<id>https://nicksxs.me/2025/12/07/%E5%9C%A8Antigravity%E4%B8%AD%E5%AF%B9%E6%AF%94%E4%BD%93%E9%AA%8CClaude-Sonnet-4-5%E6%A8%A1%E5%9E%8B/</id>
|
|
<published>2025-12-07T14:31:37.000Z</published>
|
|
<updated>2025-12-07T14:31:37.861Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前使用了Antigravity的gemini pro 3模型，发现它似乎对语义理解有点跑偏，正好再来对比下我们的常胜将军 Claude Sonnet 4.5 模型<br>我们还是用的同样的简短prompt</p>
|
|
<figure class="highlight</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>来看下我装备了5060TI显卡的gpt-oss模型表现</title>
|
|
<link href="https://nicksxs.me/2025/11/30/%E6%9D%A5%E7%9C%8B%E4%B8%8B%E6%88%91%E8%A3%85%E5%A4%87%E4%BA%865060TI%E6%98%BE%E5%8D%A1%E7%9A%84gpt-oss%E6%A8%A1%E5%9E%8B%E8%A1%A8%E7%8E%B0/"/>
|
|
<id>https://nicksxs.me/2025/11/30/%E6%9D%A5%E7%9C%8B%E4%B8%8B%E6%88%91%E8%A3%85%E5%A4%87%E4%BA%865060TI%E6%98%BE%E5%8D%A1%E7%9A%84gpt-oss%E6%A8%A1%E5%9E%8B%E8%A1%A8%E7%8E%B0/</id>
|
|
<published>2025-11-30T13:53:14.000Z</published>
|
|
<updated>2025-11-30T13:53:14.718Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前在我的3060笔记本上试了有显卡的情况下gpt-oss的表现，只能说勉强可以用，比mbp上是可用了很多，毕竟那玩意除非完全把内存都让给gpt-oss，不然都跑不起来，只是生成速度还是有点感人，差不多就4.66token&#x2F;s,一直觉得能在本地跑个稍微能用点的模型</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>来看下google最新力作Antigravity的水平如何</title>
|
|
<link href="https://nicksxs.me/2025/11/23/%E6%9D%A5%E7%9C%8B%E4%B8%8Bgoogle%E6%9C%80%E6%96%B0%E5%8A%9B%E4%BD%9Cgemini-3-pro%E7%9A%84%E6%B0%B4%E5%B9%B3%E5%A6%82%E4%BD%95/"/>
|
|
<id>https://nicksxs.me/2025/11/23/%E6%9D%A5%E7%9C%8B%E4%B8%8Bgoogle%E6%9C%80%E6%96%B0%E5%8A%9B%E4%BD%9Cgemini-3-pro%E7%9A%84%E6%B0%B4%E5%B9%B3%E5%A6%82%E4%BD%95/</id>
|
|
<published>2025-11-23T15:36:25.000Z</published>
|
|
<updated>2025-11-23T15:36:25.537Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>这是一个类似于cursor的智能ide，并且内置了最新的gemini-3-pro<br><img</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>基于iflow的kimi模型来体验代码生成</title>
|
|
<link href="https://nicksxs.me/2025/11/16/%E5%9F%BA%E4%BA%8Eiflow%E7%9A%84kimi%E6%A8%A1%E5%9E%8B%E6%9D%A5%E4%BD%93%E9%AA%8C%E4%BB%A3%E7%A0%81%E7%94%9F%E6%88%90/"/>
|
|
<id>https://nicksxs.me/2025/11/16/%E5%9F%BA%E4%BA%8Eiflow%E7%9A%84kimi%E6%A8%A1%E5%9E%8B%E6%9D%A5%E4%BD%93%E9%AA%8C%E4%BB%A3%E7%A0%81%E7%94%9F%E6%88%90/</id>
|
|
<published>2025-11-16T11:14:42.000Z</published>
|
|
<updated>2025-11-16T11:14:50.053Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>首先是生成plan<br><img data-src="https://img.nicksxs.me/uPic/TcFxPi.png"></p>
|
|
<p>完整生成完后还是无法运行，多次把问题提给kimi也没有解决</p>
|
|
<figure class="highlight</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>基于iflow的qwen模型来体验代码生成</title>
|
|
<link href="https://nicksxs.me/2025/11/09/%E5%9F%BA%E4%BA%8Eiflow%E7%9A%84qwen%E6%A8%A1%E5%9E%8B%E6%9D%A5%E4%BD%93%E9%AA%8C%E4%BB%A3%E7%A0%81%E7%94%9F%E6%88%90/"/>
|
|
<id>https://nicksxs.me/2025/11/09/%E5%9F%BA%E4%BA%8Eiflow%E7%9A%84qwen%E6%A8%A1%E5%9E%8B%E6%9D%A5%E4%BD%93%E9%AA%8C%E4%BB%A3%E7%A0%81%E7%94%9F%E6%88%90/</id>
|
|
<published>2025-11-09T13:07:39.000Z</published>
|
|
<updated>2025-11-09T13:07:39.726Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>这里我们用iflow的默认<code>Qwen3-Coder-Plus</code>模型来生成一个todo应用<br>本次的prompt稍微调整了下</p>
|
|
<figure class="highlight shell"><table><tr><td</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>体验心流助手iflow</title>
|
|
<link href="https://nicksxs.me/2025/11/02/%E4%BD%93%E9%AA%8C%E5%BF%83%E6%B5%81%E5%8A%A9%E6%89%8Biflow/"/>
|
|
<id>https://nicksxs.me/2025/11/02/%E4%BD%93%E9%AA%8C%E5%BF%83%E6%B5%81%E5%8A%A9%E6%89%8Biflow/</id>
|
|
<published>2025-11-02T10:49:41.000Z</published>
|
|
<updated>2025-11-02T10:49:41.717Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>最近一直在找这些AI编程助手，想找个cc的八分平替，刚好前阵子在朋友圈看到iflow-cli，据说是可以免费用GLM4.6这些模型，<br>首先安装也很简单<br>使用这行命令就行</p>
|
|
<figure class="highlight</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>体验Roo Code代码生成插件</title>
|
|
<link href="https://nicksxs.me/2025/10/26/%E4%BD%93%E9%AA%8CRoo-Code%E4%BB%A3%E7%A0%81%E7%94%9F%E6%88%90%E6%8F%92%E4%BB%B6/"/>
|
|
<id>https://nicksxs.me/2025/10/26/%E4%BD%93%E9%AA%8CRoo-Code%E4%BB%A3%E7%A0%81%E7%94%9F%E6%88%90%E6%8F%92%E4%BB%B6/</id>
|
|
<published>2025-10-26T13:47:18.000Z</published>
|
|
<updated>2025-10-26T13:48:03.347Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前在GLM网站看到他们家有适配各种代码生成插件，就先来体验下这个Roo Code<br>这是个在vscode插件市场里可以安装的插件，<br><img</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>在claude code中体验下GLM最新模型</title>
|
|
<link href="https://nicksxs.me/2025/10/18/%E5%9C%A8claude-code%E4%B8%AD%E4%BD%93%E9%AA%8C%E4%B8%8BGLM%E6%9C%80%E6%96%B0%E6%A8%A1%E5%9E%8B/"/>
|
|
<id>https://nicksxs.me/2025/10/18/%E5%9C%A8claude-code%E4%B8%AD%E4%BD%93%E9%AA%8C%E4%B8%8BGLM%E6%9C%80%E6%96%B0%E6%A8%A1%E5%9E%8B/</id>
|
|
<published>2025-10-18T14:33:49.000Z</published>
|
|
<updated>2025-10-18T14:33:49.405Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>在很久之前最开始体验国产模型的时候就体验了ChatGLM，那个时候属于国内基本没几家有出大语言模型，但是说实话体验效果的确比较一般<br>但是目前GLM也是国内比较头部的大语言模型了，特别是最新出的GLM4.6，看大佬的评测好像比Deepseek</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>将claude用的更有效一些的小窍门</title>
|
|
<link href="https://nicksxs.me/2025/10/12/%E5%B0%86claude%E7%94%A8%E7%9A%84%E6%9B%B4%E6%9C%89%E6%95%88%E4%B8%80%E4%BA%9B%E7%9A%84%E5%B0%8F%E7%AA%8D%E9%97%A8/"/>
|
|
<id>https://nicksxs.me/2025/10/12/%E5%B0%86claude%E7%94%A8%E7%9A%84%E6%9B%B4%E6%9C%89%E6%95%88%E4%B8%80%E4%BA%9B%E7%9A%84%E5%B0%8F%E7%AA%8D%E9%97%A8/</id>
|
|
<published>2025-10-12T12:03:15.000Z</published>
|
|
<updated>2025-10-12T12:03:15.437Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>claude的模型现在在写代码这块还是比较权威的，只是也别小瞧了本身它的通用能力<br>我们打开类似于gpt，claude经常是以文本形式的提问，再得到文本形式的回答，但是对于我们程序员来说，还有一些用途其实是可以更有效的<br>比如我们可以让claude帮我们画图<br></summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>学习下用Google的agent开发工具Agent Development Kit</title>
|
|
<link href="https://nicksxs.me/2025/10/04/%E5%AD%A6%E4%B9%A0%E4%B8%8B%E7%94%A8Google%E7%9A%84agent%E5%BC%80%E5%8F%91%E5%B7%A5%E5%85%B7Agent-Development-Kit/"/>
|
|
<id>https://nicksxs.me/2025/10/04/%E5%AD%A6%E4%B9%A0%E4%B8%8B%E7%94%A8Google%E7%9A%84agent%E5%BC%80%E5%8F%91%E5%B7%A5%E5%85%B7Agent-Development-Kit/</id>
|
|
<published>2025-10-04T14:50:43.000Z</published>
|
|
<updated>2025-10-12T12:02:33.355Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>学习下用Google的agent开发工具Agent Development Kit简称 ADK，看着名字就不一般，冲着JDK的地位去的哈哈<br>首先我们安装下pom包</p>
|
|
<figure class="highlight xml"><table><tr><td</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>怎么给wsl占用的硬盘空间做下瘦身</title>
|
|
<link href="https://nicksxs.me/2025/09/27/%E6%80%8E%E4%B9%88%E7%BB%99wsl%E5%8D%A0%E7%94%A8%E7%9A%84%E7%A1%AC%E7%9B%98%E7%A9%BA%E9%97%B4%E5%81%9A%E4%B8%8B%E7%98%A6%E8%BA%AB/"/>
|
|
<id>https://nicksxs.me/2025/09/27/%E6%80%8E%E4%B9%88%E7%BB%99wsl%E5%8D%A0%E7%94%A8%E7%9A%84%E7%A1%AC%E7%9B%98%E7%A9%BA%E9%97%B4%E5%81%9A%E4%B8%8B%E7%98%A6%E8%BA%AB/</id>
|
|
<published>2025-09-27T11:05:51.000Z</published>
|
|
<updated>2025-09-27T11:07:03.058Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>wsl出来这些年感觉在windows里使用linux，或者ubuntu真的是方便了很多，不用搞虚拟机这些<br>但是也有一些明显的问题，首先就是默认它是在C盘中安装，长久使用占用的空间就会越来越大，而且对于一些临时使用占用空间后，即使删除了文件，windows并不会对已经扩</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="wsl" scheme="https://nicksxs.me/categories/wsl/"/>
|
|
|
|
|
|
<category term="wsl" scheme="https://nicksxs.me/tags/wsl/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>通过显卡来给gpt-oss做个加速</title>
|
|
<link href="https://nicksxs.me/2025/09/21/%E9%80%9A%E8%BF%87%E6%98%BE%E5%8D%A1%E6%9D%A5%E7%BB%99gpt-oss%E5%81%9A%E4%B8%AA%E5%8A%A0%E9%80%9F/"/>
|
|
<id>https://nicksxs.me/2025/09/21/%E9%80%9A%E8%BF%87%E6%98%BE%E5%8D%A1%E6%9D%A5%E7%BB%99gpt-oss%E5%81%9A%E4%B8%AA%E5%8A%A0%E9%80%9F/</id>
|
|
<published>2025-09-21T13:58:37.000Z</published>
|
|
<updated>2025-09-21T13:58:37.755Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前在我的MacBookPro里使用gpt-oss，因为总的内存只有18g，加上系统本身和各种后台程序的占用，剩下的一般就一半左右，然后这个模型大小就有13g左右，运行起来是很吃力的<br>正好我这有个服役比较久的windows笔记本，带有3060显卡，只是显存是6g的，想</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>怎么让claude code用国内的模型</title>
|
|
<link href="https://nicksxs.me/2025/09/13/%E6%80%8E%E4%B9%88%E8%AE%A9claude-code%E7%94%A8%E5%9B%BD%E5%86%85%E7%9A%84%E6%A8%A1%E5%9E%8B/"/>
|
|
<id>https://nicksxs.me/2025/09/13/%E6%80%8E%E4%B9%88%E8%AE%A9claude-code%E7%94%A8%E5%9B%BD%E5%86%85%E7%9A%84%E6%A8%A1%E5%9E%8B/</id>
|
|
<published>2025-09-13T13:29:54.000Z</published>
|
|
<updated>2025-09-13T13:29:54.155Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前体验过gemini code这个工具，谷歌还算比较友好，有token，有网络就可以用，claude</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>用gpt-oss零改动实现一个todo应用</title>
|
|
<link href="https://nicksxs.me/2025/09/07/%E7%94%A8gpt-oss%E9%9B%B6%E6%94%B9%E5%8A%A8%E5%AE%9E%E7%8E%B0%E4%B8%80%E4%B8%AAtodo%E5%BA%94%E7%94%A8/"/>
|
|
<id>https://nicksxs.me/2025/09/07/%E7%94%A8gpt-oss%E9%9B%B6%E6%94%B9%E5%8A%A8%E5%AE%9E%E7%8E%B0%E4%B8%80%E4%B8%AAtodo%E5%BA%94%E7%94%A8/</id>
|
|
<published>2025-09-07T10:55:20.000Z</published>
|
|
<updated>2025-09-07T10:55:20.746Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><h2 id="上次是文字型的，这次我们来写点代码试试，看看效果怎么样首先肯定是比较简单的尝试开始写一个todo应用，主要是看能不能零改动就跑起来用最简单的提示词然后就给出了代码"><a</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>体验openai开源的gpt-oss模型</title>
|
|
<link href="https://nicksxs.me/2025/08/30/%E4%BD%93%E9%AA%8Copenai%E5%BC%80%E6%BA%90%E7%9A%84gpt-oss%E6%A8%A1%E5%9E%8B/"/>
|
|
<id>https://nicksxs.me/2025/08/30/%E4%BD%93%E9%AA%8Copenai%E5%BC%80%E6%BA%90%E7%9A%84gpt-oss%E6%A8%A1%E5%9E%8B/</id>
|
|
<published>2025-08-30T14:13:18.000Z</published>
|
|
<updated>2025-08-31T13:04:35.418Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>openai终于履行诺言开源了自家模型，只不过不是像deepseek那样把最新最好的开源出来，而是专门调了一大一小两个模型来开源，当然也有一些技术亮点<br>刚好我本地是18g内存的mbp，勉强可以跑20b大小的，这次开源的有20b跟200b感觉尺寸上还是比较取巧的，20b</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>真正适合spring ai的使用方式</title>
|
|
<link href="https://nicksxs.me/2025/08/24/%E7%9C%9F%E6%AD%A3%E9%80%82%E5%90%88spring-ai%E7%9A%84%E4%BD%BF%E7%94%A8%E6%96%B9%E5%BC%8F/"/>
|
|
<id>https://nicksxs.me/2025/08/24/%E7%9C%9F%E6%AD%A3%E9%80%82%E5%90%88spring-ai%E7%9A%84%E4%BD%BF%E7%94%A8%E6%96%B9%E5%BC%8F/</id>
|
|
<published>2025-08-24T15:18:00.000Z</published>
|
|
<updated>2025-08-24T15:18:00.228Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>上周被spring-ai调用模型api这个折腾的有点难受，这周又试了下用deepseek官方的api，但结果还是一样，不确定是啥原因，请求会一直卡着<br>后来想着本地有ollama，那是不是可以直接用ollama来调本地模型<br>这里我们用一个非常基础很小的模型</p></summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
<entry>
|
|
<title>差点忘了还有spring ai这个包</title>
|
|
<link href="https://nicksxs.me/2025/08/16/%E5%B7%AE%E7%82%B9%E5%BF%98%E4%BA%86%E8%BF%98%E6%9C%89springb-ai%E8%BF%99%E4%B8%AA%E5%8C%85/"/>
|
|
<id>https://nicksxs.me/2025/08/16/%E5%B7%AE%E7%82%B9%E5%BF%98%E4%BA%86%E8%BF%98%E6%9C%89springb-ai%E8%BF%99%E4%B8%AA%E5%8C%85/</id>
|
|
<published>2025-08-16T15:08:27.000Z</published>
|
|
<updated>2025-08-16T15:09:29.139Z</updated>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<summary type="html"><p>之前经常看一些LLM的工具，类似于ollama，chatbox等等，这些都类似于是个独立工具，真的在代码里使用的话可能没那么方便<br>正好看了下spring有个spring-ai的包，可以来使用试下<br>首先我们要用比较新的jdk版本，比如jdk17，然后需要使用spr</summary>
|
|
|
|
|
|
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/categories/LLM/"/>
|
|
|
|
|
|
<category term="LLM" scheme="https://nicksxs.me/tags/LLM/"/>
|
|
|
|
</entry>
|
|
|
|
</feed>
|