arxiv
PublishedJuly 21, 2026 at 4:00 AM
LaCache: Exact Caching and Precision-Adaptive Inference for Diffusion Large Language Models
Publisher summary· verbatim
arXiv:2607.16339v1 Announce Type: new Abstract: Diffusion-based Large Language Models(DLLMs) enable parallel generation via Semi-Autoregressive (SAR) decoding in text generation. However, current methods suffer from severe operator-level redundancy: they recompute the entire sequence during denoisin
Stay posted· Newsletter
A 5-min weekly brief — top movers, price watch, story of the week.
Discussion
No replies yet. Be first.
Related coverage
More from ARXIV
arxivCapacity and Redundancy Trade-offs in Multi-Task Learning5harxivPredictive Training with Latent Imagination for Visual Quadruped Navigation5harxivWhere Not to Learn: Prior-Aligned Training with Subset-based Attribution Constraints for Reliable Decision-Making5harxivSEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?5hThe Bubble Brief
WEEKLYRead AI insights every Tuesday — top movers, new releases, story of the week.
Originally published on arxiv ↗