implementation detail · filed under inference & serving
Inference-side KV-cache reset on weight synchronization
Resets the inference KV cache after a weight broadcast to prevent mixing tokens encoded under different weight versions.
Also called KV-cache reset on weight sync.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
Weight broadcasts trigger an inference-side KV-cache reset, so the cache can never mix tokens encoded under different weight versions.
usedinference servingin LAGUNA M.1/XS.2Poolside
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 replayExact SWA KV reconstruction via full multi-layer replayFine-grained prefix hashingHierarchical KV cachingKDA with prefill cacheKey-value reuse in global attention layers