MoE-ViE: Mixture of Experts Vision Encoder for Efficient Image and Video Understanding

  • 类型:arxiv
  • 标识:2608.17402
  • 链接:https://arxiv.org/abs/2608.17402
  • 主分类:multimodal
  • 形态:method
  • 被引:0
  • 被引来源:Semantic Scholar
  • S2被引:0
  • 影响力被引:0
  • TLDR:This work systematically study MoE designs for vision encoder scaling and finds that fine-grained MoE topologies yield substantial gains over both dense and standard MoE counterparts, and proposes an auxiliary-loss-free balancing variant for better expert utilization, and designs a specialized MoE kernel to mitigate inference latency overhead.
  • 待LLM分类:否
  • 标题中文:MoE-ViE:面向高效图像与视频理解的混合专家视觉编码器
  • TLDR中文:本文系统性地研究了视觉编码器扩展中的 MoE 设计,发现细粒度 MoE 拓扑相较于稠密与标准 MoE 基线均带来显著提升;提出了一种无辅助损失的均衡变体以改善专家利用率,并设计了专用 MoE kernel 以缓解推理时延开销。
  • 来源文件
  • /inbox/tom/_candidates/2026-08-20-agent-rag-longcontext-candidates.json
  • [S2 enrich]