https://txtmix.com/tags/reinforcement-learning/Recent content in Reinforcement-Learning on Text MatrixHugozh-cnTue, 21 Jul 2026 20:06:14 +0800https://txtmix.com/posts/tech/alxndr-little-book-rl-cleanrl-style-implementations/Wed, 15 Jul 2026 21:27:31 +0800https://txtmix.com/posts/tech/alxndr-little-book-rl-cleanrl-style-implementations/<h1 id="the-little-book-of-rl从零到-ppo-的-cleanrl-风格实现解析">The Little Book of RL：从零到 PPO 的 CleanRL 风格实现解析</h1> <blockquote> <p>仓库：<a href="https://github.com/alxndrTL/little-book-rl" target="_blank" rel="noopener noreffer ">alxndrTL/little-book-rl</a> 配套书 PDF：<a href="https://github.com/alxndrTL/little-book-rl/blob/main/book.pdf" target="_blank" rel="noopener noreffer ">book.pdf</a> — 21MB / V1 (June 2026) 作者：<a href="https://github.com/alxndrTL" target="_blank" rel="noopener noreffer ">Thomas Lew</a> / 471 stars / CC BY-SA 4.0 (non-commercial) 配套实现：6 个文件，~80KB，覆盖 9 个 RL 算法</p>https://txtmix.com/posts/tech/ornith-1-self-improving-agentic-coding-model-deep-dive/Tue, 30 Jun 2026 16:24:00 +0800https://txtmix.com/posts/tech/ornith-1-self-improving-agentic-coding-model-deep-dive/<h2 id="译序这份-readme-不只是模型发布说明">译序：这份 README 不只是模型发布说明</h2> <p>Ornith-1.0 是 deepreinforce-ai 团队在 2026 年发布的开源 agentic coding 模型系列。README 表面上看起来像&quot;模型发布说明&quot;，但读进去会发现它有三个不寻常的地方：</p>