2026-05-29
VideoMLA: Low-Rank Latent KV Cache for Minute-Scale Autoregressive Video Diffusion
针对视频扩散KV缓存内存开销问题，提出低秩潜在注意力机制，降低内存和延迟
Authors: Hidir Yesiltepe, Jiazhen Hu, Tuna Han Salih Meral, Adil Kaan Akan, Kaan Oktay, Hoda Eldardiry, Pinar Yanardag | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.30351v1
Abstract
Long-rollout causal video diffusion has converged on a fixed-size sliding-window KV cache, with recent progress innovating within this layout by changing which tokens occupy the window or how their positions are encoded. The per-head KV layout itself, a dominant contributor to streaming memory and latency, has been mostly left unchanged. In this paper, we present the first study of Multi-Head Latent Attention (MLA) in video diffusion. VideoMLA replaces per-head keys and values with a shared low-rank content latent and a shared decoupled 3D-RoPE positional key, reducing per-token KV memory by 92.7% at every cached layer. We further investigate why MLA succeeds in video diffusion even though the spectral assumption often used to motivate it in language models does not hold: pretrained video attention is not low-rank, with 99%-energy effective rank far above any practical latent dimension. VideoMLA retains quality at compression ratios where direct spectral approximation would predict large reconstruction error. We show that the MLA bottleneck, rather than the pretrained spectrum, determines the effective rank: both spectral and random initialization occupy nearly the full rank budget from initialization, and training preserves this budget while adapting within it. On VBench, VideoMLA matches short-horizon streaming video diffusion baselines, achieves the best overall score at long horizons among evaluated methods, and improves throughput by 1.23x on a single B200.MarginGate: Sparse Margin-Triggered Verification for Batch-Invariant LLM Inference
针对批处理LLM推理中令牌翻转问题，提出稀疏边缘触发验证方法，减少验证开销
Authors: Kexin Chu, Yang Zhou, Wei Zhang | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.30218v1
Abstract
Temperature-zero BF16 LLM inference is often treated as reproducible, yet the same request can emit different tokens when decoded alone or inside a larger batch. Existing fixes use batch-invariant operators or LLM-42's per-token verification, incurring cost even when most steps are stable. We ask whether verification can be applied exclusively to flipped tokens. Across five models, batch-induced token flips are sparse on the flip-rate benchmarks: on MATH500, Llama-3.1-8B flips on $0.48\%$ of synchronous decode steps, and all tested models stay within the 0.3-1.3% range on MATH500, GSM8K, and HumanEval. K/V perturbations remain flat before flips, while low top-1/top-2 logit margins expose much of the flip risk. MarginGate turns these observations into a verifier policy: it keeps BF16 decoding on high-margin steps, verifies only low-margin steps, and repairs confirmed mismatches by replacing the current K/V column. We evaluate on four datasets, calibrating on MATH500 and transferring to GSM8K, SharedGPT, and HumanEval. MarginGate restores 100% sequence-level deterministic decoding on Llama-3.1-8B and Qwen2.5-14B with 18.56%/15.05% verifier trigger rates, reducing LLM-42's latency increment by 2.23x/1.99x relative to always-on verification. On DSR1-Distill-Qwen-7B, the same policy reaches determinism in a harder regime at 49.50% triggers.Future Forcing: Future-aware Training-free KV Cache Policy for Autoregressive Video Generation
针对自回归视频生成KV缓存增长问题，提出未来感知无训练缓存策略，提升可扩展性
Authors: Jiayi Luo, Qiyan Liu, Tengyang Wang, JunHao Liu, Jiayu Chen, Cong Wang, Hanxin Zhu, Chen Gao, Xiaobin Hu, Qingyun Sun, Zhibo Chen | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.30083v1
Abstract
Autoregressive (AR) video generation has emerged as a promising paradigm for long-horizon video synthesis, where each frame is generated conditioned on previously generated tokens. To accelerate inference, the KV cache is used to avoid redundant recomputation across generation steps. Nevertheless, its growth with generation length introduces increasing memory and error accumulation, limiting the scalability of AR models to even longer sequences. Existing KV cache compression methods mitigate this issue by selectively retaining only video tokens deemed important. However, most existing methods assess token importance using short-horizon signals derived from the current or historical generation context, making these methods prone to overlooking tokens that appear unimportant at early steps but later become critical for future frames. In this work, we identify an important property of trained AR video models: although RoPE-modulated queries evolve across autoregressive steps, the underlying canonical pre-RoPE query distribution remains remarkably stable throughout the video generation process. This approximate stationarity implies that future query distributions are estimable from historical statistics, enabling principled future-aware cache decisions without any additional training. Building on this insight, we propose Future Forcing, a training-free future-aware KV cache policy for AR video generation. Specifically, Future Forcing first constructs a future query proxy from historical statistics, then scores KV cache tokens by their importance under this proxy, and finally merges redundant token pairs within the affine subspace induced by the future query. Extensive experiments show that Future Forcing improves long-horizon consistency under limited KV caches, achieving up to 1.49 improvement in subject consistency on VBench-Long for 60s generation over existing AR video KV cache policies.Accelerating Constrained Decoding with Token Space Compression
针对约束解码搜索空间大问题，提出令牌空间压缩方法，加速复杂语法解码
Authors: Michael Sullivan, Alexander Koller | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29986v1
Abstract
To guarantee that an LLM's outputs conform to a specified structure, context-free grammar (CFG) decoding engines force the selection of next tokens that produce strings that conform to a given CFG. While current CFG-constrained decoding engines are highly optimized, the inherent costs arising from the massive per-step search space -- i.e. the entire token vocabulary -- result in intractably high overhead for more complex CFGs: precisely the situation where CFG engines are most useful. In this paper, we introduce CFGzip, an offline technique for compressing the token search space, which massively reduces CFG engine overhead. In experiments, we report latency reduction of up to two orders of magnitude when CFGzip is used with a SoTA grammar engine, yielding an up to 7.5x speedup in total constrained generation time: with CFGzip, constrained decoding is now feasible at scale for complex CFGs.Moment-KV: Momentum-Based Decode-Time KV Cache Compression for Long Generation
针对长生成KV缓存瓶颈，提出动量解码时压缩方法，保留预填充缓存性能
Authors: Soumyadeep Jana, Sagar Nishad, Sanasam Ranbir Singh | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29873v1
Abstract
Key-Value (KV) cache remains a major bottleneck for deploying Large Language Models (LLMs) in long-generation tasks. Prior work often applies uniform compression across both prefill and decoding caches, but compressing the prefill cache degrades performance by corrupting critical context. While preserving the prefill cache is essential, decoding-phase compression remains underexplored, with existing methods relying on rigid recency windows or instantaneous attention. Our analysis of attention dynamics reveals strong temporal patterns: critical tokens receive sustained attention over long horizons, while local reasoning involves short-lived bursts. Static heuristics fail to capture this behavior, leading to premature eviction of important tokens or retention of stale ones. We propose Moment-KV, a decoding-time KV cache compression method based on momentum-driven temporal attention aggregation. Our method models token importance as a continuously evolving state, where attention is aggregated with decay, capturing both long-term influence and recent relevance. Experiments show that Moment-KV significantly improves generation fidelity in long-generation tasks (2.3-3.2 %) while maintaining decoding latency.HARP: Hadamard-Preconditioned Adaptive Rotation Processor for Extreme LLM Quantization
针对极端量化敏感问题，提出自适应旋转处理器，提升低比特量化鲁棒性
Authors: Artur Zagitov, Gleb Molodtsov, Aleksandr Beznosikov | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29843v1
Abstract
Post-training quantization (PTQ) is essential for deploying LLMs under memory and bandwidth constraints. However, extreme low-bit quantization remains highly sensitive to activation outliers and anisotropic weight curvature. Existing incoherence-based PTQ methods mitigate this issue with fixed randomized Hadamard transforms (RHTs), which improve quantization robustness but cannot adapt the rotated basis to the layer, calibration distribution, or quantizer. We introduce HARP (Hadamard-preconditioned Adaptive Rotation Processor), a learnable structured two-sided orthogonal processor that replaces fixed Hadamard mixing while preserving exact full-precision equivalence. HARP represents each rotation as a product of sparse butterfly-like block-orthogonal stages, supports non-power-of-two dimensions via Mixed-Radix schedules, and initializes to the RHT processor up to a fixed permutation. Fitted only on calibration data, HARP adapts the quantization basis to each layer and backend. Across 2-4 bit settings on models ranging from 1B to 70B parameters, HARP improves perplexity and zero-shot accuracy over fixed RHT. Importantly, HARP preserves deployment efficiency, reaching 128 tok/s versus 61 tok/s for FP16.LFQ: Logit-aware Final-block Quantization for Boosting the Generation Quality of Low-Bit Quantized LLMs
针对低比特量化生成质量下降问题，提出对数感知最终块量化，提升长响应质量
Authors: Jung Hyun Lee, June Yong Yang, Jungwook Choi, Eunho Yang | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29756v1
Abstract
As large language models continue to scale, low-bit weight-only post-training quantization (PTQ) offers a practical solution to their memory-efficient deployment. Although block-wise PTQ is capable of matching the full-precision (FP) baseline on basic language modeling and understanding, its quality is degraded for generative tasks -- especially at longer responses and extended chains of thought, which is critical in boosting task accuracy. We attribute this shortfall to two factors: (i) the omission of the unembedding layer (the LM head) in block-wise optimization and (ii) the reliance on the mean squared error (MSE) objective. Both factors cause the token probability distribution of the quantized model to misalign with that of the FP model, yielding notable accuracy drops on text generation benchmarks. To rectify the discrepancy, we introduce Logit-aware Final-block Quantization (LFQ), a simple yet effective enhancement to block-wise PTQ that quantizes the final Transformer block by minimizing the cross-entropy between the logits of the FP model and those of its quantized counterpart. By aligning token probabilities at the logit level in the final block, LFQ consistently improves the accuracy of complex generation tasks over state-of-the-art block-wise PTQ across diverse model families, while maintaining parity with FP baselines on language modeling and understanding.Domino: Decoupling Causal Modeling from Autoregressive Drafting in Speculative Decoding
针对推测解码中草稿质量与成本权衡问题，提出解耦因果建模框架，提升加速效果
Authors: Jianuo Huang, Yaojie Zhang, Qituan Zhang, Hao Lin, Hanlin Xu, Linfeng Zhang | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29707v1
Abstract
Speculative decoding accelerates LLM inference by drafting multiple tokens and verifying them in parallel with the target model. However, its practical speedup is constrained by the trade-off between draft quality and drafting cost: autoregressive drafters model causal dependencies among draft tokens but incur sequential overhead, while parallel drafters reduce drafting cost but weaken intra-block dependency modeling. In this paper, we propose Domino, a speculative decoding framework that decouples causal dependency modeling from expensive autoregressive draft execution. Domino first uses a parallel draft backbone to produce preliminary draft distributions for the entire block, and then applies a lightweight Domino head to refine them with prefix-dependent causal information. To stabilize teacher-forced causal encoding, we further introduce a base-anchored training curriculum that first strengthens the parallel backbone and then gradually shifts optimization toward the causally corrected final distribution. Experiments on Qwen3 models show that Domino achieves up to \(5.49\times\) end-to-end speedup under the Transformers backend and up to \(5.8\times\) throughput speedup under SGLang serving.AMDP: Asynchronous Multi-Directional Pipeline Parallelism for Large-Scale Models Training
针对异步流水线并行收敛问题，提出多方向异步流水线方法，维持高利用率
Authors: Ling Chen, Houming Wu, Wenjie Yu | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29664v1
Abstract
Pipeline parallelism is essential for large-scale model training, but existing asynchronous approaches often degrade convergence due to parameter mismatch between forward and backward passes. We propose Asynchronous Multi-Directional Pipeline parallelism (AMDP) to mitigate this issue while sustaining high utilization. AMDP limits the first stage of each pipeline to process at most two minibatches before backpropagation, bounding the number of parameter updates between forward and backward passes. To alleviate the resulting pipeline bubbles, AMDP launches multiple concurrent pipelines and adapts their number according to pipeline depth. In addition, AMDP accumulates gradients across minibatches and applies them in a single update, ensuring that only a bounded number of minibatches experience parameter mismatch, limited to within one optimization step. Experiments on GPT- and BERT-style models demonstrate that AMDP significantly accelerates training while preserving convergence.SAFE-Pruner: Semantic Attention-Guided Future-Aware Token Pruning for Efficient Vision-Language-Action Manipulation
针对视觉语言动作模型推理效率问题，提出语义注意力引导令牌剪枝，提升实时性
Authors: Shilin Ma, Chubin Zhang, Changyuan Wang, Yuji Wang, Yue Wu, Zixuan Wang, Jingqi Tian, Zheng Zhu, Yansong Tang | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29662v1
Abstract
Real-time inference of vision-language-action (VLA) models is essential for robotic control. While visual token pruning has shown strong potential for accelerating inference, most existing methods mainly base pruning decisions on shallow-layer cues and risk discarding visual information required by deep layers. To address this issue, we propose SAFE-Pruner, a plug-and-play pruning framework that incorporates attention cues of future layers into pruning decisions. Specifically, we identify semantic attention consistency, the tendency that VLA models concentrate their attention probability mass on the same semantic entity across execution steps. Based on this observation, we design a forward-looking strategy to forecast the token saliency in deep layers, which prevents the premature removal of critical tokens and leads to more stable acceleration. We further introduce an adaptive subtask division strategy to detect abrupt attention shifts, thereby improving forecasting accuracy and pruning reliability. Extensive experiments in simulation and real-world settings demonstrate that our method achieves up to 1.89x speedup with a minimal degradation in success rate of less than 1.7%, while outperforming state-of-the-art methods by up to 1.9%.OccamToken: Efficient VLM Inference with Training-Free and Budget-Adaptive Token Pruning
针对视觉语言模型预填充开销大问题，提出训练无关自适应令牌剪枝，提升推理效率
Authors: Geng Li, Guohao Chen, Ting Chen, Shilin Shan, Kuangji Zuo, Bofan Lyu, Tuo An, Gen Li, Jianfei Yang | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29657v1
Abstract
Vision-language models (VLMs) rely on long visual token sequences for visual understanding, making the prefill stage expensive in both computation and memory. Most existing pruning methods follow an absolute-ranking paradigm, assigning importance scores to visual tokens and retaining a fixed top-K subset. In this work, we argue that this paradigm is fundamentally brittle: attention sinks distort token importance rankings, while image redundancy and query-dependent visual evidence make fixed token budgets unreliable across inputs. We propose OccamToken, a training-free framework that replaces absolute token ranking with register-anchored relative evidence testing. Instead of asking which tokens are globally important, OccamToken evaluates whether a visual token provides information beyond a register-based reference. Our key insight is that register tokens naturally absorb low-information attention patterns, making them a stable reference for identifying genuinely informative visual evidence. Based on this principle, OccamToken performs both image-adaptive redundancy pruning and query-adaptive relevance pruning through dynamic thresholds derived from register attention. Across LLaVA-NeXT, LLaVA-v1.5, and Qwen3-VL, OccamToken consistently improves the accuracy-efficiency trade-off without additional training. Notably, on LLaVA-NeXT, it reduces 2,880 visual tokens to approximately 40 while preserving over 93% of the original accuracy, enabling stable visual token compression even in the extreme 1.4% retention regime.RTP-LLM: High-Performance Alibaba LLM Inference Engine
针对大规模LLM部署瓶颈，提出高性能推理引擎，优化加载和流水线，服务亿级用户
Authors: Boyu Tan, Jiarui Guo, Zongwei Lv, Hanbo Sun, Tong Yang, Kan Liu, Xinfei Shi, Zetao Hu, Yaxin Yu, Chi Zhang, Jianning Zhang, Xi Yang, Wei Zhang, Bo Cai, Silu Zhou, Xiyu Wang, Na He, Yinghao Yu, Wending Bao, Guiyang Huang, Yuxing Yuan, Juncheng Yin, Nan Wang, Lin Yang, Zechao Zhang, Lu Chen, Guoding Li, Tao Lan, Lin Qu | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29639v1
Abstract
Large Language Models (LLMs) have revolutionized AI applications, but deploying them at scale presents significant challenges. We present RTP-LLM, a high-performance inference engine for industrial-scale LLM deployment, successfully deployed across Alibaba Group serving over 100 million users. RTP-LLM addresses fundamental bottlenecks through integrated design. It optimizes model loading via file-order-driven I/O and parallel I/O-communication overlapping. The Prefill-Decode Disaggregation architecture decouples compute-intensive prefill from memory-bound decode phases, combined with hierarchical multi-tiered KV cache management enabling efficient cache reuse. In addition, RTP-LLM incorporates modular speculative decoding supporting multiple algorithms, adaptive KV cache quantization, and decoupled multimodal processing, with support for multi-level parallelism. Comprehensive evaluations across diverse model architectures (8B-235B parameters) have been conducted, where both controlled benchmarks and real production workloads are used. The results demonstrate RTP-LLM's superior performance against vLLM and SGLang: 4.7x-6.3x model loading speedup, 35-37% TTFT P95 latency reduction with 215% cache reuse improvement in production traffic scheduling, 1.12x-2.48x and 1.86x-2.52x throughput improvements in speculative decoding and multimodal inference, respectively, and 35-40% batch latency reduction with 1.9x-3.0x TTFT improvement in quantized inference. RTP-LLM's production-proven architecture and open-source availability make it a comprehensive solution for industrial LLM deployment.AsymVLM: Asymmetric Token Pruning for Efficient Vision-Language Model Inference
针对视觉语言模型推理效率问题，提出不对称令牌剪枝，区分视觉和文本模态
Authors: Yilin Feng, Ahmed Burak Gulhan, Mahmut Taylan Kandemir | Date: 2026-05-28
Link: http://arxiv.org/abs/2605.29535v1