https://blog.fupengcheng.top/authors/fupcode/南冥鹏影Hugo 0.168.0 & FixIt v1.0.0-beta.4zh-CNThu, 15 Oct 2026 17:51:49 +0800https://blog.fupengcheng.top/posts/research-notes/paper-review-redial/Thu, 17 Sep 2026 17:00:00 +0800https://blog.fupengcheng.top/posts/research-notes/paper-review-redial/科研笔记通过单点突变观察 embedding 的变化，直接提取 RNA 语言模型中的结构耦合信号，并用逐层分析和消融检验模型冗余。https://blog.fupengcheng.top/posts/devops/systemd-uvicorn-fastapi/Sat, 05 Sep 2026 18:19:11 +0800https://blog.fupengcheng.top/posts/devops/systemd-uvicorn-fastapi/开发运维使用 systemd 管理 Uvicorn 运行 FastAPI，配置开机启动、异常重启、日志查看和 NGINX 反向代理。https://blog.fupengcheng.top/posts/egpu/egpu-audio-problem/Tue, 25 Aug 2026 16:04:48 +0800https://blog.fupengcheng.top/posts/egpu/egpu-audio-problem/外接显卡本文记录一次真实故障：Windows 电脑通过 OCuLink 接入 NVIDIA eGPU 后，开机偶尔不显示 Wi-Fi、音量和系统托盘。最终通过禁用 <code>NVIDIA High Definition Audio</code> 解决。https://blog.fupengcheng.top/posts/egpu/wallpaper-engine-smart-start/Fri, 21 Aug 2026 18:00:33 +0800https://blog.fupengcheng.top/posts/egpu/wallpaper-engine-smart-start/外接显卡<!--post--> <h2 class="heading-element" id="背景"><span>背景</span> <a href="#%e8%83%8c%e6%99%af" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>家里使用外接显卡（eGPU）连接大屏幕时，希望 Wallpaper Engine 自动启动显示动态壁纸。</p>https://blog.fupengcheng.top/posts/algorithm/perf-dict-with-heap/Thu, 06 Aug 2026 13:15:48 +0800https://blog.fupengcheng.top/posts/algorithm/perf-dict-with-heap/算法<!--post--> <h2 class="heading-element" id="问题模式"><span>问题模式</span> <a href="#%e9%97%ae%e9%a2%98%e6%a8%a1%e5%bc%8f" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>维护一个动态字典：</p>https://blog.fupengcheng.top/posts/network/problem-tun-github-error/Thu, 21 May 2026 22:01:56 +0800https://blog.fupengcheng.top/posts/network/problem-tun-github-error/网络<!--post--> <h2 class="heading-element" id="问题"><span>问题</span> <a href="#%e9%97%ae%e9%a2%98" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>开启代理软件的 TUN 模式后，进行 <code>git push</code> 会失败，显示</p>https://blog.fupengcheng.top/posts/devops/problem-docker-timezone/Tue, 19 May 2026 17:38:41 +0800https://blog.fupengcheng.top/posts/devops/problem-docker-timezone/开发运维<p>最近将微信小程序后端迁移到 Docker 后，出现了“每日刷新”相关 Bug，例如：</p> <ul> <li>每日广告次数未刷新</li> <li>次日仍然沿用前一天状态</li> </ul> <p>排查后发现，问题并不在业务逻辑，而是：</p>https://blog.fupengcheng.top/posts/devops/certbot-webroot/Tue, 19 May 2026 03:04:28 +0800https://blog.fupengcheng.top/posts/devops/certbot-webroot/开发运维<p>之前我一直使用：</p> <pre><code>certbot --nginx</code></pre><p>来自动申请和配置 HTTPS 证书。</p> <p>这种方法虽然方便，但在实际使用中，我遇到了比较严重的问题：</p> <ul> <li>certbot 会自动修改 nginx 配置</li> <li>certbot 会自动 reload/restart nginx，导致systemd 状态与 nginx 实际状态脱节，出现“游离 nginx”，然后第一个证书成功、后面失败</li> </ul> <p>最终决定迁移到：</p>https://blog.fupengcheng.top/posts/research-notes/paper-review-trrna2/Thu, 30 Apr 2026 17:05:46 +0800https://blog.fupengcheng.top/posts/research-notes/paper-review-trrna2/科研笔记<p>trRosettaRNA2 利用二级结构 (SS) 做了两件很有意思的工作</p> <ol> <li>让SS加权于attention</li> <li>通过多 SS 条件输入实现 RNA 构象分布</li> </ol> <h2 class="heading-element" id="ss加权于attention"><span>SS加权于attention</span> <a href="#ss%e5%8a%a0%e6%9d%83%e4%ba%8eattention" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p><img loading="lazy" src='https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs42256-026-01223-x/MediaObjects/42256_2026_1223_Fig1_HTML.png?as=webp' alt="ss"></p>https://blog.fupengcheng.top/posts/devops/problem-docker-volume-edit-bug/Thu, 12 Mar 2026 19:11:13 +0800https://blog.fupengcheng.top/posts/devops/problem-docker-volume-edit-bug/开发运维<h2 class="heading-element" id="问题背景"><span>问题背景</span> <a href="#%e9%97%ae%e9%a2%98%e8%83%8c%e6%99%af" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>Docker Compose 里经常会把单个配置文件 bind mount 到容器中：</p>https://blog.fupengcheng.top/posts/devops/vps-configure/Thu, 26 Feb 2026 17:36:17 +0800https://blog.fupengcheng.top/posts/devops/vps-configure/开发运维<h2 class="heading-element" id="一系统配置"><span>一、系统配置</span> <a href="#%e4%b8%80%e7%b3%bb%e7%bb%9f%e9%85%8d%e7%bd%ae" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><h3 class="heading-element" id="更新系统"><span>更新系统</span> <a href="#%e6%9b%b4%e6%96%b0%e7%b3%bb%e7%bb%9f" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h3><pre><code>apt update &amp;&amp; apt full-upgrade -y &amp;&amp; apt autoremove -y &amp;&amp; apt autoclean -y</code></pre><h3 class="heading-element" id="安装常用工具"><span>安装常用工具</span> <a href="#%e5%ae%89%e8%a3%85%e5%b8%b8%e7%94%a8%e5%b7%a5%e5%85%b7" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h3><pre><code>apt install sudo htop git wget curl vim tmux net-tools unattended-upgrades fail2ban -y</code></pre><h3 class="heading-element" id="配置自动更新策略"><span>配置自动更新策略</span> <a href="#%e9%85%8d%e7%bd%ae%e8%87%aa%e5%8a%a8%e6%9b%b4%e6%96%b0%e7%ad%96%e7%95%a5" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h3><p>编辑文件：</p>https://blog.fupengcheng.top/posts/environment/multimolecule-low-glibc/Thu, 08 Jan 2026 13:29:03 +0800https://blog.fupengcheng.top/posts/environment/multimolecule-low-glibc/环境配置<h2 class="heading-element" id="背景"><span>背景</span> <a href="#%e8%83%8c%e6%99%af" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>我们课题组计算集群使用的glib版本是2.17，这导致了一些较新的科学计算库在安装或使用时出现问题。查看glibc版本方法：</p>https://blog.fupengcheng.top/posts/devops/problem-wx307/Sat, 29 Nov 2025 06:42:35 +0800https://blog.fupengcheng.top/posts/devops/problem-wx307/开发运维<p>最近在将自研小程序从测试版发布到正式版时，遇到了一个非常奇怪的坑：接口在本地开发和测试环境中一切正常，但一旦发布到正式版，就出现了 <strong>code <code>600002</code></strong> 的报错，提示 <strong>URL not in domain list</strong>。</p>https://blog.fupengcheng.top/posts/network/problem-custom-host/Sat, 29 Nov 2025 06:42:14 +0800https://blog.fupengcheng.top/posts/network/problem-custom-host/网络<h2 class="heading-element" id="问题背景"><span>问题背景</span> <a href="#%e9%97%ae%e9%a2%98%e8%83%8c%e6%99%af" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>使用家用Wi-Fi部署一些服务后，要访问的话，打ip地址太长了，让不懂web相关知识的人手打地址更是灾难。因此，路由器自定义host是一个好的方案。</p>https://blog.fupengcheng.top/posts/deep-learning/lora/Thu, 20 Nov 2025 19:05:47 +0800https://blog.fupengcheng.top/posts/deep-learning/lora/深度学习<p>在深度学习微调大模型时，<strong>LoRA（Low-Rank Adaptation）</strong> 提供了一种高效、轻量的微调方法。本文讲述了LoRA的原理，同时结合一个小型 MNIST 实验做了验证。</p>https://blog.fupengcheng.top/posts/research-notes/paper-review-rhofold/Sat, 15 Nov 2025 17:07:04 +0800https://blog.fupengcheng.top/posts/research-notes/paper-review-rhofold/科研笔记<h2 class="heading-element" id="模型架构"><span>模型架构</span> <a href="#%e6%a8%a1%e5%9e%8b%e6%9e%b6%e6%9e%84" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>RhoFold+是一个全可微分的端到端框架，从RNA序列输入直接输出全原子3D结构。架构分为几个关键模块：</p>https://blog.fupengcheng.top/posts/environment/dl-build/Sat, 15 Nov 2025 00:29:34 +0800https://blog.fupengcheng.top/posts/environment/dl-build/环境配置<h2 class="heading-element" id="深度学习环境搭建anacondacuda130cudnn"><span>深度学习环境搭建（Anaconda、CUDA13.0、cuDNN）</span> <a href="#%e6%b7%b1%e5%ba%a6%e5%ad%a6%e4%b9%a0%e7%8e%af%e5%a2%83%e6%90%ad%e5%bb%baanacondacuda130cudnn" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>如果不是从源码构建，那么Pytorch和OpenMM安装不需要CUDA13.0和cuDNN。</p>https://blog.fupengcheng.top/posts/environment/conda-dependence-upgrade/Fri, 14 Nov 2025 22:51:34 +0800https://blog.fupengcheng.top/posts/environment/conda-dependence-upgrade/环境配置<h2 class="heading-element" id="问题背景"><span>问题背景</span> <a href="#%e9%97%ae%e9%a2%98%e8%83%8c%e6%99%af" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>Conda、Docker 等技术的确非常强大，它们可以实现“一次开发，到处部署”，在操作系统层面保证兼容性。然而，当遇到 <strong>硬件不兼容</strong> 时，这类工具就无能为力，需要手动干预。</p>https://blog.fupengcheng.top/posts/deep-learning/cnn-params/Fri, 07 Nov 2025 00:40:31 +0800https://blog.fupengcheng.top/posts/deep-learning/cnn-params/深度学习<h2 class="heading-element" id="卷积层convolutional-layer-conv-k-n"><span>卷积层（Convolutional Layer, CONV-K-N）</span> <a href="#%e5%8d%b7%e7%a7%af%e5%b1%82convolutional-layer-conv-k-n" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p><strong>功能</strong>：提取局部特征。</p>https://blog.fupengcheng.top/posts/deep-learning/mlp-manual/Fri, 07 Nov 2025 00:34:24 +0800https://blog.fupengcheng.top/posts/deep-learning/mlp-manual/深度学习<h2 class="heading-element" id="一数据集的下载和处理"><span>一、数据集的下载和处理</span> <a href="#%e4%b8%80%e6%95%b0%e6%8d%ae%e9%9b%86%e7%9a%84%e4%b8%8b%e8%bd%bd%e5%92%8c%e5%a4%84%e7%90%86" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><h3 class="heading-element" id="下载数据"><span>下载数据</span> <a href="#%e4%b8%8b%e8%bd%bd%e6%95%b0%e6%8d%ae" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h3><p>直接调用官方库</p>https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse5/Fri, 07 Nov 2025 00:08:32 +0800https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse5/逆向渗透<p><strong>远程命令注入</strong>（Remote Command Injection，简称 RCE）是一类严重的安全漏洞，通常出现在 Web 应用或网络服务中。它允许攻击者向目标服务器发送恶意命令，从而在服务器上执行任意代码。对渗透测试者而言，远程命令注入不仅提供了直接控制目标系统的途径，还能帮助进一步扩展攻击面，提升权限或植入持久后门。</p>https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse4/Fri, 07 Nov 2025 00:03:25 +0800https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse4/逆向渗透<h2 class="heading-element" id="一反弹shell"><span>一、反弹Shell</span> <a href="#%e4%b8%80%e5%8f%8d%e5%bc%b9shell" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p>渗透测试中，如果发现命令执行漏洞，通常会希望获得一个交互式 shell。如果无法添加新账户、SSH 密钥或 <code>.rhosts</code> 文件来直接登录，下一步通常是尝试建立反向 shell 或绑定 shell 到一个 TCP 端口。</p>https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse3/Thu, 06 Nov 2025 23:46:13 +0800https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse3/逆向渗透<h2 class="heading-element" id="一sqlmap简介"><span>一、sqlmap简介</span> <a href="#%e4%b8%80sqlmap%e7%ae%80%e4%bb%8b" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><p><code>sqlmap</code> 是一个强大的开源工具，用于自动化检测和利用SQL注入漏洞。它支持多种数据库类型，包括MySQL、Oracle、PostgreSQL、SQL Server等。<code>sqlmap</code>能够帮助渗透测试人员快速检测注入点、获取数据库结构、甚至直接获取数据库中的敏感数据。</p>https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse2/Thu, 06 Nov 2025 02:21:38 +0800https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse2/逆向渗透<h2 class="heading-element" id="一metasploit"><span>一、Metasploit</span> <a href="#%e4%b8%80metasploit" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><h3 class="heading-element" id="msf简介"><span>MSF简介</span> <a href="#msf%e7%ae%80%e4%bb%8b" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h3><p>Metasploit Framework(MSF)是一款开源安全漏洞检测工具，附带数千个已知的软件漏洞，并保持持续更新。Metasploit可以用来信息收集、漏洞探测、漏洞利用等渗透测试的全流程。</p>https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse1/Thu, 06 Nov 2025 00:29:26 +0800https://blog.fupengcheng.top/posts/reverse-engineering/learning-reverse1/逆向渗透<h2 class="heading-element" id="一主机发现"><span>一、主机发现</span> <a href="#%e4%b8%80%e4%b8%bb%e6%9c%ba%e5%8f%91%e7%8e%b0" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h2><h3 class="heading-element" id="arp-scan与arping"><span>arp-scan与arping</span> <a href="#arp-scan%e4%b8%8earping" class="heading-mark"><svg class="octicon octicon-link" viewBox="0 0 16 16" version="1.1" width="16" height="16" aria-hidden="true"><path d="m7.775 3.275 1.25-1.25a3.5 3.5 0 1 1 4.95 4.95l-2.5 2.5a3.5 3.5 0 0 1-4.95 0 .751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018 1.998 1.998 0 0 0 2.83 0l2.5-2.5a2.002 2.002 0 0 0-2.83-2.83l-1.25 1.25a.751.751 0 0 1-1.042-.018.751.751 0 0 1-.018-1.042Zm-4.69 9.64a1.998 1.998 0 0 0 2.83 0l1.25-1.25a.751.751 0 0 1 1.042.018.751.751 0 0 1 .018 1.042l-1.25 1.25a3.5 3.5 0 1 1-4.95-4.95l2.5-2.5a3.5 3.5 0 0 1 4.95 0 .751.751 0 0 1-.018 1.042.751.751 0 0 1-1.042.018 1.998 1.998 0 0 0-2.83 0l-2.5 2.5a1.998 1.998 0 0 0 0 2.83Z"></path></svg></a> </h3><p>arp-scan 可用于快速发现同一局域网下的主机：</p>