<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Yestino - The Signal · Skip SDPA</title><link>https://yestino.com/zh-CN/entities/skip-sdpa-47bebc</link><description>与 Skip SDPA 相关的全部事件</description><language>zh-CN</language><atom:link href="https://yestino.com/zh-CN/entities/skip-sdpa-47bebc/feed.xml" rel="self" type="application/rss+xml"/><item><title>trunk/610343e5acab74781f4ac043633d32465f8a86e5: [inductor] Skip SDPA fusion when the matched scale is a Tensor (#191496)</title><link>https://yestino.com/zh-CN/events/trunk-610343e5acab74781f4ac043633d32465f8a86e5-inductor-skip-d897b8</link><guid isPermaLink="true">https://yestino.com/zh-CN/events/trunk-610343e5acab74781f4ac043633d32465f8a86e5-inductor-skip-d897b8</guid><pubDate>Wed, 26 Aug 2026 14:28:42 GMT</pubDate><description>Fixes #191203 What happens Writing attention as q @ k.transpose(-2, -1) / scale with a 0-d tensor scale works in eager but crashes under torch.compile: Runtime…
来源：PyTorch Releases</description></item></channel></rss>