specific method · filed under inference & serving
Exact SWA KV reconstruction via full multi-layer replay
Reconstructs SWA KV states exactly by replaying the recent window across all layers.
- source
- 1
- model
- 1
- labs adopt it
- 0
- strongest
- not used
How sources treat it
One count per evidence span, weakest treatment to strongest.
not used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
Exact reconstruction requires replaying the most recent L×n win tokens, where L is the number of layers and nwin is the SWA window size.
not usedinference servingin DeepSeek-V4DeepSeek
Filed alongside
Other methods under inference & serving :: KV cache management.
Chunked prefillPrefix cachingPrompt cachingSWA Bounded ReplayLanguage-model-only serving modePeriodic cache checkpointingCompressed KV cachingKDA-aware prefix-cache managementRadixCacheZero SWA caching8-bit Mamba cache quantizationAsynchronous cache offloading and restorationAutomatic cacheBlock-based KV cacheCross-layer KV-cache reuseCustomized heterogeneous KV-cache layoutDisable prefix caching for benchmarkingDynamic allocation of fixed-size state-cache poolsEncoder SWA bounded replayFine-grained prefix hashingHierarchical KV cachingInference-side KV-cache reset on weight synchronizationKDA with prefill cacheKey-value reuse in global attention layers