<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Yestino - The Signal · Fuse NVFP4</title><link>https://yestino.com/zh-CN/entities/fuse-nvfp4-c6dd12</link><description>与 Fuse NVFP4 相关的全部事件</description><language>zh-CN</language><atom:link href="https://yestino.com/zh-CN/entities/fuse-nvfp4-c6dd12/feed.xml" rel="self" type="application/rss+xml"/><item><title>trunk/d25057513dc155ee32bfdf4e760da11652269423: [inductor][NVGEMM] Fuse NVFP4 output scaling (#194651)</title><link>https://yestino.com/zh-CN/events/trunk-d25057513dc155ee32bfdf4e760da11652269423-inductor-nvge-254883</link><guid isPermaLink="true">https://yestino.com/zh-CN/events/trunk-d25057513dc155ee32bfdf4e760da11652269423-inductor-nvge-254883</guid><pubDate>Sat, 03 Oct 2026 01:41:51 GMT</pubDate><description>Human commentary: ModelOpt NVFP4 linear layers multiply _scaled_mm output by a scalar dequantization factor.
来源：PyTorch Releases</description></item></channel></rss>