41 articles tagged with #kv-cache. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Researchers propose EinSort, an adaptive tensorization method that uses index ordering to identify and compress low-rank structures in large language models, demonstrating improved results for weight and KV-cache compression compared to existing approaches.
Researchers introduce AGENTSERVESIM, a hardware-aware simulator designed to evaluate serving policies for multi-turn LLM agents without requiring expensive accelerator deployments. The simulator accurately reproduces real-system performance within 6% error while running on standard CPUs, enabling scalable exploration of agent-serving policies across different hardware configurations and workload scenarios.
Researchers introduce SemanticSeg, a large semantic segmentation dataset, and block distillation framework to improve block attention mechanisms for long-context language models. The approach uses a frozen full-attention teacher to train block-attention students more efficiently, addressing key challenges in KV cache reuse for applications like RAG.
Researchers introduce Murmur, an inference system that optimizes long-form automatic speech recognition by balancing accuracy and latency through a two-level approach: intermediate chunk sizes at the inter-chunk level and attention sparsity exploitation at the intra-chunk level. The system achieves 4.2x latency reduction while maintaining single-pass accuracy on benchmark tests.
Researchers introduce COVER, a new verification technique for diffusion language models that eliminates inefficient token oscillations during parallel decoding. By using KV cache overrides to preserve context while selectively verifying tokens in a single forward pass, COVER accelerates inference while maintaining output quality.
Researchers propose block-based double decoders, a transformer architecture that combines the training efficiency of decoder-only models with the inference speed advantages of encoder-decoder models. The innovation uses doubly-causal block-based attention masks to enable full loss supervision and static sequence packing, achieving 2/3 reduction in KV-cache memory and per-token compute at inference time.
Researchers propose Shadow Mask Distillation to address the memory bottleneck created by KV cache compression during reinforcement learning post-training of large language models. The technique tackles the critical off-policy bias that emerges when compressed contexts are used during rollout generation while full contexts are used for parameter updates, a problem that amplifies instability in RL optimization.
Fluxion, a new hybrid CPU-GPU system, optimizes long-context inference by efficiently managing key-value caches split between host and GPU memory. The approach delivers 1.5x-3.7x speedup over existing baselines while maintaining near-baseline accuracy, addressing a critical bottleneck in modern large language model deployment.
Researchers propose a novel self-indexing KV cache system that unifies compression and retrieval for efficient sparse attention in large language models. The method uses 1-bit vector quantization and integrates with FlashAttention to reduce memory bottlenecks in long-context LLM inference.
Researchers propose OxyGen, a unified KV cache management system for Vision-Language-Action Models that enables efficient multi-task parallelism in embodied AI agents. The system achieves up to 3.7x speedup by sharing computational resources across tasks and eliminating redundant processing of shared observations.
Researchers have developed LookaheadKV, a new framework that significantly improves memory efficiency in large language models by intelligently evicting less important cached data. The method achieves superior accuracy while reducing computational costs by up to 14.5x compared to existing approaches, making long-context AI tasks more practical.
OrbitFlow is a new KV cache management system for long-context LLM serving that uses adaptive memory allocation and fine-grained optimization to improve performance. The system achieves up to 66% better SLO attainment and 3.3x higher throughput by dynamically managing GPU memory usage during token generation.
Researchers have introduced PiKV, an open-source KV cache management framework designed to optimize memory and communication costs for Mixture of Experts (MoE) language models across multi-GPU and multi-node inference. The system uses expert-sharded storage, intelligent routing, adaptive scheduling, and compression to improve efficiency in large-scale AI model deployment.
Researchers from PKU-SEC-Lab have developed KEEP, a new memory management system that significantly improves the efficiency of AI-powered embodied planning by optimizing KV cache usage. The system achieves 2.68x speedup compared to text-based memory methods while maintaining accuracy, addressing a key bottleneck in memory-augmented Large Language Models for complex planning tasks.
Researchers introduce SideQuest, a novel KV cache management system that uses Large Reasoning Models to compress memory usage during long-horizon AI tasks. The system reduces peak token usage by up to 65% while maintaining accuracy by having the model itself determine which tokens are useful to keep in memory.
The article discusses the implementation of KV (Key-Value) cache mechanisms in nanoVLM, a lightweight vision-language model framework. This technical implementation focuses on optimizing memory usage and inference speed for multimodal AI applications.
