<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>语言模型架构 on AI内参</title>
    <link>https://www.neican.ai/main_topics/%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B%E6%9E%B6%E6%9E%84/</link>
    <description>Recent content in 语言模型架构 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Mon, 16 Jun 2025 12:30:04 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/main_topics/%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B%E6%9E%B6%E6%9E%84/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>超越Transformer：混合扩散模型Eso-LM以65倍速重塑语言生成范式</title>
      <link>https://www.neican.ai/insights/transformereso-lm65-20250616123004/</link>
      <pubDate>Mon, 16 Jun 2025 12:30:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/transformereso-lm65-20250616123004/</guid>
      <description>康奈尔和CMU研究者推出了名为Eso-LM的新型语言模型，它将离散扩散模型与自回归模型相结合，实现了推理速度高达65倍的突破，同时提升了生成质量并克服了传统扩散模型的效率瓶颈。这项创新通过引入KV缓存、灵活的注意力机制和混合训练策略，有望重塑大语言模型的架构格局，并在AI业界引起了英伟达、谷歌等巨头的关注，预示着语言生成技术迈向更高效、更实用的新阶段。</description>
    </item>
  </channel>
</rss>
