<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[Qwen Flash Next for AMD R9700]]></title><description><![CDATA[<p dir="auto"><a href="https://github.com/lcchan0525sg/Qwen-Flash-Next-Strata-for-AMD-R9700" rel="nofollow ugc">https://github.com/lcchan0525sg/Qwen-Flash-Next-Strata-for-AMD-R9700</a></p>
<p dir="auto">适用于 AMD R9700 的 Qwen Flash Next（基于 Strata）；包含安装指南、性能调优、基准测试及第三方声明的源码分发包。</p>
<p dir="auto">Radeon AI PRO R9700、Ryzen 9 5900X、128 GiB 内存、IQ3_S 量化、配置 128K 上下文长度，且 CPU 视觉功能处于空闲状态。输出速率 45 to 56 token/s</p>
<p dir="auto">欢迎评论</p>
]]></description><link>https://lcz.me/topic/2235</link><generator>RSS for Node</generator><lastBuildDate>Mon, 12 Oct 2026 12:24:28 GMT</lastBuildDate><atom:link href="https://lcz.me/topic/2235.rss" rel="self" type="application/rss+xml"/><pubDate>Sat, 10 Oct 2026 01:50:40 GMT</pubDate><ttl>60</ttl><item><title><![CDATA[Reply to Qwen Flash Next for AMD R9700 on Sat, 10 Oct 2026 06:44:53 GMT]]></title><description><![CDATA[<p dir="auto">QFN 精确编码可能不如27B。但是一般的代码编写我感觉足够用了。<br />
另外我用它来识别报表图片（主要内容是员工手写的产品数据，还有员工姓名），总体感比27B又快又准。</p>
]]></description><link>https://lcz.me/post/23727</link><guid isPermaLink="true">https://lcz.me/post/23727</guid><dc:creator><![CDATA[stxpnet]]></dc:creator><pubDate>Sat, 10 Oct 2026 06:44:53 GMT</pubDate></item><item><title><![CDATA[Reply to Qwen Flash Next for AMD R9700 on Sat, 10 Oct 2026 06:39:52 GMT]]></title><description><![CDATA[<blockquote>
<p dir="auto"><a class="plugin-mentions-user plugin-mentions-a" href="/user/exe127" aria-label="Profile: exe127">@<bdi>exe127</bdi></a> <a href="/post/23716">said</a>:</p>
<p dir="auto">125B輸出的結果真的會比 27B更強嗎?</p>
</blockquote>
<p dir="auto">我也非常好奇這一點 兩者還沒對決完畢</p>
<p dir="auto">不過可以先參考大家的使用體驗 (每個人使用的量化程度不同 框架不同 可能感受也會不同) :</p>
<blockquote>
<p dir="auto"><strong>體感經驗 哪一個強？？ Qwen3.8-Flash vs. Qwen3.8-27B</strong><br />
<a href="https://lcz.me/post/22008">https://lcz.me/post/22008</a></p>
</blockquote>
<p dir="auto">目前來說 確定Flash 比較強的是 "美感" 和 顯卡溫度比較低</p>
]]></description><link>https://lcz.me/post/23725</link><guid isPermaLink="true">https://lcz.me/post/23725</guid><dc:creator><![CDATA[kos or]]></dc:creator><pubDate>Sat, 10 Oct 2026 06:39:52 GMT</pubDate></item><item><title><![CDATA[Reply to Qwen Flash Next for AMD R9700 on Sat, 10 Oct 2026 06:34:41 GMT]]></title><description><![CDATA[<blockquote>
<p dir="auto"><a class="plugin-mentions-user plugin-mentions-a" href="/user/exe127" aria-label="Profile: exe127">@<bdi>exe127</bdi></a> <a href="/post/23716">说</a>:</p>
<p dir="auto"><a class="plugin-mentions-user plugin-mentions-a" href="/user/kos-or" aria-label="Profile: kos-or">@<bdi>kos-or</bdi></a> strata 用的125B輸出的結果真的會比 27B更強嗎?</p>
</blockquote>
<p dir="auto">Q3_S 现在日常办公感觉比以前的MXFP4 要强，毕竟模型体量在这里。</p>
]]></description><link>https://lcz.me/post/23719</link><guid isPermaLink="true">https://lcz.me/post/23719</guid><dc:creator><![CDATA[张光璞]]></dc:creator><pubDate>Sat, 10 Oct 2026 06:34:41 GMT</pubDate></item><item><title><![CDATA[Reply to Qwen Flash Next for AMD R9700 on Sat, 10 Oct 2026 06:31:47 GMT]]></title><description><![CDATA[<p dir="auto"><a class="plugin-mentions-user plugin-mentions-a" href="/user/kos-or" aria-label="Profile: kos-or">@<bdi>kos-or</bdi></a> strata 用的125B輸出的結果真的會比 27B更強嗎?</p>
]]></description><link>https://lcz.me/post/23716</link><guid isPermaLink="true">https://lcz.me/post/23716</guid><dc:creator><![CDATA[exe127]]></dc:creator><pubDate>Sat, 10 Oct 2026 06:31:47 GMT</pubDate></item><item><title><![CDATA[Reply to Qwen Flash Next for AMD R9700 on Sat, 10 Oct 2026 02:30:36 GMT]]></title><description><![CDATA[<p dir="auto">Strata QFN 對於上下文长度好像沒什麼限制 只要内存足夠</p>
<blockquote>
<p dir="auto">"Qwen3.8-Flash-Next 原生最大 Context 為 262,144 tokens（256K），透過 YaRN 可擴展至 1,000,000 tokens（1M）。"</p>
</blockquote>
]]></description><link>https://lcz.me/post/23664</link><guid isPermaLink="true">https://lcz.me/post/23664</guid><dc:creator><![CDATA[kos or]]></dc:creator><pubDate>Sat, 10 Oct 2026 02:30:36 GMT</pubDate></item><item><title><![CDATA[Reply to Qwen Flash Next for AMD R9700 on Sat, 10 Oct 2026 02:19:42 GMT]]></title><description><![CDATA[<p dir="auto"><a href="https://github.com/Niko1221/Strata" rel="nofollow ugc">https://github.com/Niko1221/Strata</a><br />
原项目已经支持 AMD 显卡，我的项目只是在此基础上进行了一些调整优化。</p>
]]></description><link>https://lcz.me/post/23663</link><guid isPermaLink="true">https://lcz.me/post/23663</guid><dc:creator><![CDATA[LC LC]]></dc:creator><pubDate>Sat, 10 Oct 2026 02:19:42 GMT</pubDate></item></channel></rss>