<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>LoRA on AI内参</title>
    <link>https://www.neican.ai/tags/lora/</link>
    <description>Recent content in LoRA on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Mon, 20 Oct 2025 11:10:05 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/lora/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>OPPO AndesVL：开源多模态大模型如何重塑端侧AI，驱动下一代AI手机的“即时智能”革命</title>
      <link>https://www.neican.ai/insights/oppo-andesvlaiai-20251020111005197-3/</link>
      <pubDate>Mon, 20 Oct 2025 11:10:05 +0800</pubDate>
      <guid>https://www.neican.ai/insights/oppo-andesvlaiai-20251020111005197-3/</guid>
      <description>OPPO推出的AndesVL开源多模态大模型，以其0.6B-4B的灵活尺寸、极致的端侧优化（如稀疏化、QALFT、OKV）和在多项基准测试中SOTA的表现，重新定义了AI手机的性能与隐私界限。这一战略性开源不仅强化了OPPO的市场竞争力，更通过软硬件协同和开放生态，加速了行业向高性能、高隐私、低延迟的“即时智能”体验迈进，预示着手机AI将从云端走向真正以用户为中心的个人化新时代。</description>
    </item>
    <item>
      <title>边缘智能的突破：小米小爱同学如何在资源受限下实现高性能大模型推理</title>
      <link>https://www.neican.ai/insights/article-20250624231007315-0/</link>
      <pubDate>Tue, 24 Jun 2025 23:10:07 +0800</pubDate>
      <guid>https://www.neican.ai/insights/article-20250624231007315-0/</guid>
      <description>小米小爱同学团队在端侧大模型部署方面取得了显著进展，通过自研推理框架、动态优化、投机推理、量化以及创新的“共享基座+LoRA”架构，成功克服了移动设备资源限制，实现了高性能、多任务并发。文章深入剖析了小米的技术策略，并展望了未来硬件与模型架构（如Linear Attention）在推动端侧AI普惠化中的关键作用。</description>
    </item>
  </channel>
</rss>
