<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0">
    <channel>
        <title>微调 - 标签 - Simi Studio</title>
        <link>/tags/%E5%BE%AE%E8%B0%83/</link>
        <description>微调 - 标签 - Simi Studio</description>
        <generator>Hugo -- gohugo.io</generator><language>zh-CN</language><managingEditor>simi@simi.studio (Simi)</managingEditor>
            <webMaster>simi@simi.studio (Simi)</webMaster><lastBuildDate>Thu, 25 Dec 2025 21:28:00 &#43;0800</lastBuildDate><atom:link href="/tags/%E5%BE%AE%E8%B0%83/" rel="self" type="application/rss+xml" /><item>
    <title>LLM 微调入生产：QLoRA 与 RLHF 实战</title>
    <link>/posts/llm-finetuning-production/</link>
    <pubDate>Thu, 25 Dec 2025 21:28:00 &#43;0800</pubDate>
    <author>simi@simi.studio (Simi)</author>
    <guid>/posts/llm-finetuning-production/</guid>
    <description><![CDATA[QLoRA + RLHF 是把 LLM 微调用于生产的主流方案。这篇文章快速过一遍实战流程。]]></description>
</item>
<item>
    <title>本地微调 LLM：ollama &#43; unsloth 实战经验</title>
    <link>/posts/local-llm-finetuning-guide/</link>
    <pubDate>Wed, 24 Dec 2025 10:45:00 &#43;0800</pubDate>
    <author>simi@simi.studio (Simi)</author>
    <guid>/posts/local-llm-finetuning-guide/</guid>
    <description><![CDATA[开源微调工具越来越成熟，在本地用消费级 GPU 微调一个小模型已经可行。这篇文章是实战经验：用什么工具、数据准备、常见坑，以及什么场景值得微调。]]></description>
</item>
</channel>
</rss>
