FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention 文章

ArXiv CS.AI2026-06-09NEWSen作者: Yan Wang, Qifan Zhang, Jiachen Yu, Tian Liang, Dongyang Ma, Xiang Hu, Zibo Lin, Chunyang Li, Zhichao Wang, Jia Li, Yujiu Yang, Haitao Mi, Dong Yu

查看原文 →

FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention · 相关技术

相关技术

dual-encoder architecture massive backbone model standard retrieval training frameworks standard dual-encoder architecture backbone-free decoupled training strategy DeepSeek-V4 Neural Memory Indexer Lookahead Sparse Attention