efficient-long-context-reasoning

【效率与协作】Achieve 10× higher decoding throughput on long prompts by replacing 50% of cross-attention layers with gated memory units (GMUs) combining SSMs and attention. Maintains reasoning capability while reducing memory I/O bottleneck from O(d_kv·N) to O(d_h).

独立 GitHub 快照 · 自动分类 · 以原作者文档为准