<?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[Strata Qwen3.8-flash-next-iq2_xs + AMD RX9070XT 16gb VRAM]]></title><description><![CDATA[<p dir="auto">各位大神大家好,</p>
<p dir="auto">本來買這張顯卡只是為了打遊戲，後來認識本地AI一直在試Qwen 3.5 27b Q4 等等不同的模型，如果強行用qwen3.8 的量化最後也只能4至5tk/s, 最近認識strata 後，也放上來給大夥看看效果。因為工作關係，所以測試的就是叫Anything LLM 在PDF讀取後製作一個mermaid workflow.</p>
<p dir="auto">硬件如下:<br />
CPU - AMD Ryzen 9 5950X — 16C / 32T<br />
Motherboard - ASUS TUF Gaming B550M-PLUS<br />
RAM -64 GB DDR4<br />
GPU - AMD Radeon RX 9070 XT 16 GB — Sapphire OC<br />
OS	Windows 11<br />
AI stack	 Strata + AnythingLLM</p>
<p dir="auto">對比之前Qwen 3.8 27b Q4, 4-5 tk/sec, 效果算是很不錯<br />
<img src="https://upload.lcz.me/uploads/4bb74cf6-20cb-414f-b791-b09278701d5d.png" alt="Screenshot 2026-10-06 100730.png" class="img-markdown  img-fluid" /></p>
]]></description><link>https://lcz.me/topic/2127</link><generator>RSS for Node</generator><lastBuildDate>Mon, 12 Oct 2026 19:42:49 GMT</lastBuildDate><atom:link href="https://lcz.me/topic/2127.rss" rel="self" type="application/rss+xml"/><pubDate>Tue, 06 Oct 2026 02:16:17 GMT</pubDate><ttl>60</ttl><item><title><![CDATA[Reply to Strata Qwen3.8-flash-next-iq2_xs + AMD RX9070XT 16gb VRAM on Tue, 06 Oct 2026 02:51:03 GMT]]></title><description><![CDATA[<p dir="auto">看來有可能Zen3架構跟DDR4稍微拖慢了一點 目前看到都是Zen5特別快 建議STRATA版本列出 並且Prefill測速一下</p>
]]></description><link>https://lcz.me/post/22723</link><guid isPermaLink="true">https://lcz.me/post/22723</guid><dc:creator><![CDATA[dardeaw feng]]></dc:creator><pubDate>Tue, 06 Oct 2026 02:51:03 GMT</pubDate></item></channel></rss>