标签
HarmoCore 引入了一种功能性潜扩散方法,用于从稀疏传感器数据中重建振荡波场,在二维和三维场景中以最小感知实现了效率和精度的显著提升。
This paper introduces S4R, a low-rank KV cache compression method that combines selective token sampling, subspace construction, and sparse reconstruction to achieve up to 5× compression with near full-cache accuracy on LongBench and RULER benchmarks.