[Paper & Models] MMFlash: Faster Speculative Decoding for Vision-Language Models

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1xxmd3w/paper_models_mmflash_faster_speculative_decoding/
🏷️ 分组: Reddit|AI·自托管·VPS
🕒 时间: 2026-11-16 18:48:24 From the LocalLLaMA community on Reddit: [Paper & Models] MMFlash: Faster Speculative Decoding for Vision-Language Models