Efficient VLM Inference

VLM: Vision-Language Model

Follow this topic Ask about this topic

Momentum

38 papers in the last four weeks, up 100% on the four weeks before. 0.4% of all new papers.

Jul 13Week of Sep 28

Latest papers 322

Aug 2, 2026cs.RO

VespaSeg: A Resource-Aware Ground-then-Segment Pipeline for Referring Expression Segmentation

Referring expression segmentation requires language conditioned localization and pixel-accurate masks, but monolithic models can be costly to deploy. We present VespaSeg, a modular pipeline that grounds a text query with a compact vision-language model and converts the predicted box to a mask with MobileSAM. We study Florence-2-base, Florence-2-large, and Moondream2 grounders together with targeted adaptation of the grounding and segmentation stages. Under a repository-specific RefCOCO validation protocol containing the first expression for each of 3,811 referenced-object records, the adapted Florence-2-base pipeline obtains 73.64 mean intersection over union (mIoU) and 84.60 precision at IoU 0.5. On an NVIDIA RTX 6000 Ada GPU it processes 22.8 cached-image queries per second with 2.20 GB mean allocated GPU memory. A matched 500-query comparison gives 73.73 mIoU for Florence-2-base and 72.82 for Florence-2-large, while the base model is 1.70 times faster and uses 1.17 GB less allocated memory. Ablations show that ground-truth-box adaptation raises MobileSAM mIoU from 82.22 to 86.61 and that reducing the Florence-2 output-token budget from 64 to 32 preserves accuracy. These results support compact, modular grounding and segmentation, while also exposing the need for evaluation on the complete standard RefCOCO expression splits and deployment hardware.

Savindu Dilshan Wickramasinghe

University of Moratuwa · Department of Electronic and Telecommunication Engineering, University of Moratuwa Moratuwa, Sri Lanka

Promptable Image Segmentation Efficient VLM Inference Referring Expression Segmentation Referring Image Segmentation Vision-Language Grounding

Aug 2, 2026cs.CV

MonitorVLM-v2: A Deployed Vision-Language Framework for Real-Time Safety Violation Detection

Large vision--language models (VLMs) can reason step by step about complex visual scenes, but this open-ended, autoregressive chain-of-thought (CoT) approach is poorly suited to safety-critical, rule-governed settings such as industrial surveillance, where decisions must be bounded, deterministic, and low-latency. Because CoT inference cost scales jointly with reasoning length and the number of concurrent streams, it creates a throughput bottleneck that precludes the real-time, multistream monitoring required for industrial accountability. Here we present MonitorVLM-v2, a deployment-oriented framework that recasts VLM-based safety assessment as probabilistic inference over a finite regulatory decision space, compressing multimodal reasoning into single-step rule-ID predictions and reducing decoding from a variable-length sequence to a single token. We introduce symbolic policy optimization (SymPO), a novel contrastive policy optimization algorithm that sharpens decision boundaries within this finite symbolic space, together with an entropy-driven triage mechanism that routes uncertain predictions to human reviewers for expert confirmation. In a four-month prospective deployment across 10 concurrent camera feeds in an operational underground mining facility, MonitorVLM-v2 achieved a 19.45-fold increase in inference speed and identified 2.78 times as many confirmed violations as the site's routine manual inspection workflow, demonstrating the practical value of compressed symbolic decision-making for real-time, auditable industrial monitoring.

Jiang Wu, Sichao Wu, Yinsong Ma +2

School of Mechanical Engineering, University of Science and Technology Beijing · The Laboratory for Computational Sensing and Robotics, Johns Hopkins University

Vision-Language Models Efficient VLM Inference Policy Optimization Video Surveillance

Aug 2, 2026cs.RO

FreqNav: Stage-Wise Frequency Routing for Object-Oriented Aerial Vision-Language Navigation

Object-oriented aerial vision-and-language navigation (VLN) requires searching for a described target and landing on it precisely, under long-horizon and closed-loop control. Guided by a target-descriptive instruction during navigation, perceptual priorities dynamically evolve: early-stage exploration prioritizes low-frequency spatial layout, and then shifts to high-frequency target details. Existing VLN methods model the varying perceptual requirements across navigation stages with identical visual tokens, leading to interference from irrelevant objects and background clutter. To this end, we therefore formulate long-horizon aerial navigation as a frequencypreference shift from spatial structure to local detail and propose FreqNav, a lightweight frequency-routing adaptive perception framework. Under a fixed computational budget, FreqNav dynamically reallocates visual tokens across frequency components according to the current navigation stage. A Frequency Token Router selects stage-relevant visual representations from dual-view observations, while a Phase-dependent Grounding Module anchors visual evidence through explicit supervision. A Diffusion Transformer then predicts smooth trajectories for continuous control. Experiments show that FreqNav outperforms strong baselines while achieving approximately 3x faster inference. Real-world deployment further demonstrates its effectiveness, efficiency, and practical potential for long-horizon aerial autonomy.

Yin Tang, Jiawei Ma, Jiahao Li +4

Big Data Institute, Central South University · Department of Computer Science, City University of Hong Kong · School of Computer Science and Engineering, Central South University +1

Efficient VLM Inference UAV Navigation Vision-Language Navigation Aerial VLN

Aug 1, 2026cs.CV

Coverage-Driven Adaptive Keyframe Selection for Video Understanding

Recent advances in large vision-language models (LVLMs) have enabled long-video understanding and analysis. However, processing the large number of frames in a video incurs substantial computational overhead. Existing methods reduce LVLM inference costs by scoring frame-query relevance before inference and selecting keyframes accordingly. Nevertheless, the distribution of relevant frames varies across queries, and these methods often need to score hundreds or thousands of frames. To address this limitation, we propose CSES, a training-free semantic keyframe selector that adaptively determines the numbers of frames to score and keyframes to select. CSES estimates the prominence of the frame-query relevance profile to guide active acquisition and adapt the temporal coverage of each input. It then formulates keyframe selection as a coverage problem that jointly accounts for semantic relevance, temporal redundancy, and visual redundancy. Active acquisition and keyframe selection terminate based on coverage saturation. The selection objective is monotone and submodular, enabling greedy optimization with a standard approximation guarantee. Experiments with four LVLMs on two benchmarks show that our method preserves accuracy while scoring

4

-

13\times

fewer frames and selecting

18.4\%

-

20.5\%

fewer input keyframes than existing baselines. CSES further achieves a

3.1

-

5.4\times

speedup in frame selection over baselines.

Junyang Zhang, Puhan Luo, Chen Tang +2

University of Science and Technology of China, Hefei, China

Efficient VLM Inference Large Vision-Language Models Video Frame Selection Query-Aware Video Frame Selection

Jul 31, 2026cs.CV

Latent-Centroid Steering: Single-Pass Classifier-Free Guidance for Command-Aligned Autonomous Driving

Vision-language models (VLMs) have recently emerged as a promising paradigm for end-to-end autonomous driving, enabling agents to map multimodal inputs and high-level navigation instructions directly to executable trajectories. However, in practice, these models exhibit a persistent command-following gap: predicted trajectories often show weak sensitivity to navigation commands, resulting in incorrect behavior at critical decision points. We identify this issue as a form of conditional policy collapse, where regression-based training under multimodal trajectory distributions encourages the model to rely on dominant visual priors while marginalizing the language-conditioned signal. To address this issue, we introduce a principled formulation of classifier-free guidance (CFG) for regression-based vision-language driving. We show that CFG can be interpreted as isolating the instruction-induced residual in the action space by contrasting conditional and unconditional predictions, thereby explicitly amplifying the effect of the navigation command at inference time. However, a standard two-pass CFG introduces prohibitive latency for real-time control and produces noisy instance-level guidance directions. Building on a mean-shift interpretation of CFG, we propose Latent-Centroid Steering (LCS), a single-pass guidance mechanism that replaces instance-level residuals with class-level latent shifts. By projecting conditional representations toward precomputed command-specific centroids, LCS performs class-level latent steering based on cluster geometry that is both more stable and computationally efficient. We demonstrate that LCS reduces inference latency by approximately 50% while achieving stronger command adherence and improved driving performance on both closed-loop (Bench2Drive) and open-loop (nuScenes) benchmarks. Code are available at https://github.com/codingmlinprocess/LCS.

Meibo Hu, Jiamian Wang, Pichao Wang +1

Rochester Institute of Technology · NVIDIA

Classifier-Free Guidance VLMs for Autonomous Driving End-to-End Autonomous Driving Efficient VLM Inference Autonomous Driving

Jul 30, 2026cs.CV

LAST: The Last Query Token Guides Visual Token Pruning for Edge-Cloud Collaborative MLLM Inference

Multimodal foundation models are reshaping edge-cloud visual intelligence from task-specific feature pipelines into token-based interfaces, where edge devices encode visual inputs into tokens for a general-purpose cloud MLLM. However, dense visual-token sequences increase cloud-side inference costs. Existing pruning methods mainly target centralized inference: vision-driven methods can operate before cloud execution but are typically query-agnostic, whereas query-guided methods often rely on internal states of the target MLLM and cannot determine token relevance before transmission. Compact guidance models offer an alternative, but existing designs may require costly attention aggregation or auxiliary generation. We propose LAST, a training-free framework for query-dependent visual token pruning in edge-cloud collaborative MLLM inference. LAST uses a compact edge-side VLM as a guidance proxy and derives a lightweight importance signal from the last query token's attention to visual tokens. Under causal attention, the last query token can attend to the full visual sequence and the entire query context, enabling query-aware pruning without cloud-model access, autoregressive generation, or costly aggregation over multiple query positions. LAST then retains a diverse set of query-relevant visual tokens under a fixed token budget. We evaluate LAST on 11 multimodal benchmarks under multiple token budgets against pruning methods with different guidance strategies. Experiments show that LAST consistently achieves the strongest performance, preserving 95.4% of the full-token accuracy while retaining only 12.5% of the visual tokens, with low edge-side selection overhead and reduced cloud-side computation.

Feng Yang, Xinrui Ju, Keyang Zhang +6

1City University of Hong Kong · 2Peng Cheng Laboratory

Efficient VLM Inference Visual Token Pruning Edge-Cloud Computing

Jul 30, 2026cs.CV

Thinking Once Is Enough: Intermediate-Layer Evidence Routing for High-Resolution VQA

High-resolution visual question answering (HR-VQA) is often treated as a problem of insufficient evidence acquisition, where failing multimodal large language models must inspect images again through cropping, re-encoding, or multi-round search. We show that this view is incomplete: in many cases, fine-grained evidence has already survived visual encoding and become identifiable and influential within an intermediate-layer routing window, but is later diluted before answer generation. We propose Thinking-Once, a \textbf{training-free, single-visual-pass} evidence-routing method that reconstructs question-conditioned attention at this window, preserves core entity tokens and compact background context, and routes this evidence to later layers without extra visual encoding. Across five base models, Thinking-Once consistently improves or matches the corresponding base setting, increasing the average scores on V

^*

Bench, HRBench-4K, and HRBench-8K by \textit{+3.1}, \textit{+3.0}, and \textit{+2.7} points while reducing the average peak memory by about 4,GB. On Qwen2.5-VL-7B, it improves the three benchmarks by \textit{+9.9}, \textit{+4.6}, and \textit{+5.5} points, raising the cross-benchmark mean from 72.5 to 79.1. With the ZwZ-8B base model, Thinking-Once reaches a mean score of 82.7. Against 11 open-source HR-VQA baselines, it obtains the best or tied-best score on all three benchmark averages and the best overall mean; for example, compared with DeepScan, it reduces V

^*

Bench inference time by \textbf{97.2%} while improving the cross-benchmark mean from 77.8 to 79.1. These results show that HR-VQA can be improved by routing already encoded evidence rather than repeatedly acquiring new visual inputs. Code is available in the appendix.

Zhongkuan Mao, Xianjie Liu, Tianyu Meng +9

National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University

Visual Question Answering Efficient VLM Inference

Jul 30, 2026cs.CV

Calibrate Before Reason: Robust Visual Token Reduction against Semantic Drift in VLMs

Large Vision-Language Models (VLMs) suffer from prohibitive inference overhead due to long sequences of visual tokens. However, existing visual token reduction methods mainly improve efficiency by pruning or compressing redundant tokens without examining whether the resulting representation remains semantically consistent with the original representation. Mapping the original N-token visual sequence to K tokens may discard, dilute, or misassign critical visual cues, triggering severe semantic drift that deviates the VLM's understanding. In this paper, we first introduce the principle of 'Calibrate Before Reason' to visual token reduction and propose CaRe, a training-free robust framework that calibrates compact visual representations before reasoning to preserve semantic fidelity in VLMs. CaRe consists of two mutually complementary modules: 1) Perturbation-Robust Calibration Anchoring, which identifies calibration anchors with stable model-side influence under multi-directional perturbations; 2) Confidence-Gated Token Calibration, which extracts reliable calibration signals from unselected tokens and injects them into anchors. Extensive evaluations across diverse VLM architectures and benchmarks verify that CaRe outperforms state-of-the-art token reduction baselines. While pruning 94.4% of visual tokens, our method retains 96.4% of the original full-token performance, delivering up to 2.30 times faster end-to-end inference speed relative to unpruned vanilla models.

Jiasheng Li, Zhong Ji, Yan Zhang +1

Tianjin University, Tianjin, China · Tsinghua University, Beijing, China

VLM Robustness Efficient VLM Inference Visual Token Compression

Jul 29, 2026cs.CV

Beyond Accuracy: Auditing Spatial Provenance in Visual Token Pruning for OCR-Critical MLLM Inference

Visual-token pruning is usually judged by answer quality at a fixed retention budget. For text-rich multimodal large language models (MLLMs), this protocol can miss a distinct failure: an answer remains correct even when no retained token is locally traceable to the small OCR region that supports it. We turn this blind spot into an evidence-risk audit that couples answer behavior with geometric token-origin provenance, interventions, and realized cost; transparent training-free selectors isolate controlled operating points. On locked image-disjoint confirmation, Qwen Target at 30% retention has observed accuracy 0.786 versus 0.783 for Full (paired image-cluster difference +0.003, 95% CI [-0.014, +0.020]), yet same-budget Target, Random, and Grid retain sharply different positive-support coverage: 0.620, 0.270, and 0.318. Across Qwen3-VL-8B, LLaVA-1.5-7B, and InternVL3.5-8B, matched controls, interventions, detector tests, and external methods reveal model-specific quality-risk-traceability frontiers that accuracy alone does not expose. Materialized prefixes yield up to 4.32x batch-prefill speedup and 76.4% lower incremental peak memory; full-validation TextVQA and DocVQA further show that favorable target-verification points do not imply task-general compression. Visual-token pruning should therefore report surviving spatial provenance and realized cost alongside quality and compression.

Feixiang Liu, Qiang Qiu, Hao Zhang +1

State Key Laboratory of AI Safety, Institute of Computing Technology, CAS · University of Chinese Academy of Sciences

VLM Evaluation Efficient VLM Inference Visual Token Pruning

Jul 29, 2026cs.CV

MedARC: Training-Free Adaptive Redundancy Compression of Visual Tokens for 3D Medical Vision-Language Models

Integrating 3D medical images with vision-language models (VLMs) holds substantial promise for computer-aided diagnosis. However, volumetric images generate prohibitively long visual-token sequences with considerable spatial and inter-slice redundancy. Existing token compression methods typically apply uniform reduction or rely on a single importance signal, increasing the risk of removing regions that are clinically relevant to the query or structurally distinctive. To address this limitation, we propose MedARC, a unified, training-free framework for Adaptive Redundancy Compression of visual tokens in 3D medical VLMs. MedARC estimates token importance by integrating three complementary cues: self-attention from the VLM vision encoder, which reflects the model's intrinsic visual focus; similarity between projected visual tokens and text embeddings, which identifies query-relevant regions; and deviations of local visual foundation model features from the volume-level feature center, which highlight structurally distinctive anatomy. The resulting importance distribution guides a saliency-aware merging strategy that preserves informative tokens while consolidating redundant ones rather than simply discarding them. Experiments on CT-RATE and MR-RATE show that MedARC reduces visual-token overhead and inference time while preserving or improving diagnostic performance. Its multi-cue scoring cost is outweighed by the savings from processing fewer tokens, with greater benefits expected for larger language models.

Yitao Zhu, Mengjun Liu, Yingji Fu +2

Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong · School of Integrated Circuits and Electronics, Beijing Institute of Technology, China · 3Mental Health Research Center, The Hong Kong Polytechnic University, Hong Kong +1

Efficient VLM Inference Medical VLMs Multimodal Token Compression Visual Token Compression 3D VLMs

Jul 27, 2026cs.CV

Mage-VL: An Efficient Codec-Native Streaming Multimodal Foundation Model

Standard vision-language models (VLMs) suffer from Moravec's paradox: they excel at complex offline visual reasoning but struggle with simple streaming perception tasks and process them inefficiently. We present Mage-VL, an efficient codec-native streaming foundation model for real-time multimodal understanding and interaction. At its core, our custom tokenizer, Mage-ViT, replaces uniform frame sampling by selectively encoding dynamic, entropy-rich regions using motion vectors and residual energy across sparse anchor (I) and predicted (P) frames. Operating at a 16 x 16 patch level, this reduces visual token consumption by over 75% while preserving spatiotemporal context. Trained from scratch on approximately 560M unlabeled images and 100M unlabeled video frames, Mage-ViT matches or outperforms flagship encoders trained on billions of image-text pairs. We establish AI4AI data pipelines encompassing prompt-code joint optimization for multimodal captioning and AI-driven performance diagnosis to guide training recipes. Furthermore, through a bio-inspired dual-system architecture - a lightweight System 1 event gate and a causal System 2 decoder - Mage-VL enables proactive streaming perception. Extensive evaluations show that Mage-VL-4B matches Qwen3-VL-4B on static tasks while achieving strong gains in video understanding and 2D/3D spatial reasoning, with up to a 3.5x wall-clock inference speedup, and comprehensively surpasses the 15B Phi-4-reasoning-vision baseline. Beyond model artifacts, we deliver seven key empirical findings covering pre-training data efficiency, variable-resolution scaling, codec system acceleration, VideoQA SFT redundancy, motion-spatial synergy, AI4AI data pipelines, and Zero-Vision SFT for multimodal RL.

Senqiao Yang, Kaichen Zhang, Zhaoyang Jia +20

Microsoft Mage Team

Vision-Language Models Efficient VLM Inference Visual Tokenization Video-Language Models

Jul 27, 2026cs.CV

CADER: Confidence-Aware Dynamic Evidence Reasoning for Long-Video Understanding

Long-video understanding increasingly relies on large vision-language models and tool-augmented reasoning, but most systems apply the same inference procedure to every example regardless of difficulty. This uniform strategy invokes unnecessary tool-assisted processing for easy questions and provides limited control when difficult questions require fine-grained temporal evidence. We propose CADER (Confidence-Aware Dynamic Evidence Reasoning), a training-free framework for adaptive and reliable long-video reasoning. CADER first performs global reasoning over uniformly sampled frames and estimates answer confidence with a logit-margin signal, allowing high-confidence examples to exit early. For uncertain examples, CADER activates a second-stage tool-augmented loop that combines temporal cropping, lightweight semantic verification, and Relevance-Guided Resampling to progressively localize question-relevant evidence. This design treats tool use as a sample-level decision: a single global pass handles easy cases, while additional reasoning is reserved for examples where uncertainty suggests that more evidence is needed. Experiments on multiple VideoQA benchmarks show that CADER improves long-video reasoning while bypassing Stage~2 for high-confidence samples. Moreover, when applied to a backbone trained only with tool-free chain-of-thought supervision, CADER achieves competitive performance against specialized tool-augmented frameworks, suggesting a practical inference-time route for adaptive long-video reasoning.

Jinlong Yang, Wenhao Zhang, Kuanwei Lin +1

School of Computer Science, Northwestern Polytechnical University · School of Electronic and Computer Engineering, Peking University · Department of Computer Science and Technology, Tsinghua University

Tool Use in VLMs Efficient VLM Inference Long-Video Understanding Video Reasoning Long-Video QA

Jul 27, 2026cs.CV

MAViE: A Multi-scale Adaptive Vision Encoder for Fine-grained Visual Perception and Efficient Multimodal Reasoning

Vision-language models commonly project all tokens produced by a pretrained vision encoder into a large language model. However, final-layer features can discard text, local attributes, and spatial relationships, while high-resolution inputs substantially increase context length and inference latency. We introduce \method, a Multi-scale Adaptive Vision Encoder. \method uses position-dependent gates to fuse shallow, intermediate, and deep features from a vision Transformer, preserving global semantics while enhancing edges, text, and local structure. It then performs question-conditioned token routing according to question relevance, local information content, global semantics, and spatial coverage, with a token budget that adapts to image complexity. To mitigate compression loss, we further introduce full-to-compressed representation distillation and a spatial diversity regularizer. In an illustrative simulation under a unified 7B language-model framework, \method reduces the average number of SigLIP-SO400M visual tokens from 729 to 146 (approximately 80.0%) and improves the mean score on VQAv2, GQA, TextVQA, ScienceQA-IMG, and MMBench by 2.2 percentage points, while reducing single-image time to first token from 228,ms to 129,ms. We provide the full model design and evaluation protocol. All reported numbers currently serve only as placeholders for paper organization and experimental design; formal claims require real training runs, independent replications, and official benchmark evaluation.

Shaofei Lei

Efficient VLM Inference Multi-Scale Feature Fusion VLM Reasoning Visual Token Compression

Jul 27, 2026cs.AI

GOTS: Greedy Orthogonal Token Selection for High-Resolution Vision-Language Models

Modern vision-language models (VLMs) increasingly rely on dynamic or high-resolution visual encoding, producing thousands of visual tokens that substantially increase downstream language-model inference cost. Existing token-reduction methods assess token utility through token-wise importance, query relevance, coverage, pairwise diversity, or subset-level objectives. Our key insight is to view visual token reduction through selected-span complementarity: instead of scoring a token in isolation or through pairwise relations, we assess how much of its feature is orthogonal to the span of the already retained subset. Based on this perspective, we propose Greedy Orthogonal Token Selection (GOTS), a training-free and query-agnostic method. At each step, GOTS selects the token with the largest residual energy orthogonal to the current retained span. This rule exactly maximizes the one-step augmented Gram determinant among candidate additions, giving each greedy step a precise local geometric guarantee for subset expansion. Across five high-resolution VLM backbones from the Qwen-VL and InternVL families and eleven diverse benchmarks, GOTS achieves higher average performance retention than the strongest evaluated baselines, and a controlled OCRBench study shows that it reduces model-side time-to-first-token after accounting for selection overhead. Code is available at https://github.com/newLLing/GOTS.

Jun Ling, Tao Huang, Junzhuo Liu +2

School of Computer Science and Engineering, University of Electronic Science and Technology of China

Efficient VLM Inference Large Vision-Language Models Visual Token Pruning Visual Token Compression

Jul 26, 2026cs.CV

PathSelect: Sequential Token Selection for Whole Slide Pathology

Gigapixel Whole-Slide Images (WSIs) present a fundamental computational bottleneck for vision-language models (VLMs) due to extreme sequence lengths. Existing approaches predominantly rely on spatial sampling or training-free pruning, which risk diluting weak but informative signals, leading to the loss of critical diagnostic evidence due to the spatially diffuse nature of pathological cues. We reformulate WSI token pruning as a sequential selection process, enabling the model to autonomously learn an optimal routing strategy rather than relying on static heuristics. We herein propose a decoupled routing framework integrated as an active plugin into the fully pre-trained SlideChat base model, leaving both the slide encoder and large language model frozen. To provide continuous gradients for the non-differentiable pruning operation during training, we introduce PathSelect. PathSelect employs a variance-preserving noise gate to modulate each patch's information flow via a differentiable Soft Top-K operator, paired with a diagonal-attention Denoiser that recovers the perturbed representations without semantic leakage. At inference, the PathSelect module is entirely detached. Relying solely on the trained Scorer, a deterministic Hard Top-K operator executes adaptive, data-dependent trajectory termination, significantly accelerating downstream generative processing with exceptionally low sequential token selection latency. Driven by an empirical average of only 44.86 tokens under a maximum constraint of K = 128, our framework achieves 74.00% overall accuracy on SlideBench (TCGA), representing an approximate 36.6x spatial token reduction relative to the uncompressed baseline average while consistently outperforming sampling-based counterparts.

Jingzhi Chen, Landi He, Zehong Chen +2

Shenzhen University of Advanced Technology

Efficient VLM Inference Medical VLMs Visual Token Pruning Whole-Slide Image Analysis

Jul 26, 2026cs.CV

MemVLN: Episodic and Procedural Memory for Vision-and-Language Navigation

Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires agents to maintain long-horizon visual history for trajectory consistency while executing actions with low latency. Existing video-based VLN approaches typically struggle to satisfy both demands simultaneously. To address these challenges, we propose MemVLN, a novel VLN framework that achieves state-of-the-art performance with real-time inference efficiency (14 FPS). MemVLN utilizes a visual encoder to process continuous observations and a Large Language Model (LLM) to interpret instructions and generate actions. Central to our approach is an Episodic Memory management that applies pyramidal resolutions. This mechanism concentrates computation on immediate percepts while retaining compressed long-term history. Complementing to this design, we introduce Procedural Memory for fast action with a compact vocabulary of atomic mid-level actions to bypass auto-regressive decoding latency. Experiments on VLN-CE show that MemVLN-4B surpasses the baseline Qwen3-VL-4B architecture by 5.8% SR in R2R and 9.7% SR in RxR, while achieving a 7

\times

speedup in inference latency.

Yuqi Liu, Shengju Qian, Tianyuan Qu +5

1The Chinese University of Hong Kong · 2LIGHTSPEED · 3The University of Hong Kong +1

Memory-Augmented VLMs Efficient VLM Inference Efficient VLA Models Vision-Language Navigation

Jul 25, 2026cs.CV

UltraViT: Latency-Optimized On-device Vision Encoder for Large Vision-Language Models

Large Vision-Language Models (LVLMs) remain bottlenecked by massive computational footprints, precluding their deployment on resource-constrained edge devices. While efforts to compress LVLMs focus heavily on vision token reduction or smaller language models, the vision encoder is largely overlooked, typically deployed as a monolithic, computationally heavy feature extractor. Moreover, there is no previous effort that designs a vision encoder for LVLMs directly optimized for on-device latency. In this paper, we present UltraViT, a vision encoder for LVLMs, explicitly designed and optimized for on-device performance. Specifically, by taking into account real on-device latencies, we systematically design a pyramidal architecture that strategically integrates and adapts heterogeneous spatial mixers at the macro-block level. Furthermore, to pre-train UltraViT, we propose a novel two-stage generative pre-training strategy: cultivating rich spatial features via dense distillation, followed by direct generative supervision from a capacity-mixed frozen LLM. Compared to standard contrastive and SSL, we show that our pre-training is much more effective for achieving high-level semantic grounding for UltraViT needed for the subsequent generative multimodal alignment of LVLM training. Extensive experiments demonstrate that our on-device latency-informed design combined with our tailored training strategy establishes a new state-of-the-art for efficient LVLM encoding, significantly outperforming existing encoder-centric baselines while operating on-device at nearly 1.7xthe speed.

Ioannis Maniadis Metaxas, Adrian Bulat, Alberto Baldrati +4

Samsung AI Cambridge · Technical University of Iasi · Queen Mary University of London

Vision-Language Models Efficient VLM Inference Efficient ViTs Vision-Language Pretraining

Jul 25, 2026cs.CV

WaveZip: Wavelet-Driven Space-Time Decoupling for Video Token Condensation

Existing Large Vision-Language Models (LVLMs) struggle with long-form video understanding due to the quadratic computational cost of visual tokens. While recent efficient methods attempt to compress tokens via hard pruning or uniform merging, they operate strictly in the spatial feature domain, where robust structural context and discriminative semantic details are inherently entangled. In this work, we propose WaveZip, a joint signal-frequency-domain framework for efficient video inference. Driven by the insight that temporal redundancy resides in low-pass approximation scales while spatial saliency strongly correlates with high-frequency components, WaveZip leverages Discrete Wavelet Transforms (DWT) to disentangle these signals. Temporally, it employs 1D DWT to analyze query-frame relevance, and the resulting high-frequency coefficients are further gated by inter-frame differences, with both signals jointly driving the dynamic allocation of a precise frame-level token budget. Spatially, a 2D DWT decomposes features into low-frequency approximations and high-frequency detail components, where the high-frequency coefficients are modulated within query-salient regions to regulate spatial reconstruction. Importantly, WaveZip requires no task-specific training and can be seamlessly integrated into off-the-shelf LVLMs to boost inference efficiency. Extensive experiments on long video understanding benchmarks demonstrate that WaveZip retains 99.6% of the full performance under an extreme 10x compression ratio, consistently outperforming state-of-the-art methods.

Yuhui Zeng, Wang Chen, Jinfa Huang +5

Media Analytics and Computing Lab, Xiamen University, Xiamen, China · Department of Computer Science, University of Rochester, Rochester, NY, USA

Efficient VLM Inference Video Token Compression Visual Token Compression Temporal Video Understanding

Jul 25, 2026cs.CV

Structured Redundancy Modeling for Efficient Visual Token Pruning in High-Resolution MLLMs

Recent high-resolution Multimodal Large Language Models (MLLMs) generate thousands of visual tokens per input, leading to a visual token explosion that introduces severe latency bottlenecks. While token pruning mitigates this issue, state-of-the-art subset-optimization methods typically rely on iterative subset construction to jointly capture visual diversity and instruction relevance. As visual token counts scale, this sequential dependency introduces significant selection overhead, severely limiting the translation of theoretical FLOPs reductions into actual wall-clock speedups. To address this limitation, we propose Single-Forward Pruner (SFPruner), a structural reformulation of visual token pruning that embeds redundancy control directly into the scoring space, bypassing the need for iterative combinatorial optimization. Our non-iterative framework achieves redundancy-aware importance selection in a single forward pass through two complementary mechanisms. First, to attenuate redundancy at the covariance level, we introduce a semantics-guided ridge leverage scheme. By integrating instruction relevance and visual saliency, this mechanism suppresses dominant covariance directions and mitigates representation bias. Second, ranking-based directional masking resolves residual overlap through asymmetric similarity competition, where higher-scoring tokens explicitly suppress redundant lower-scoring alternatives via parallel tensor operations. Extensive evaluations demonstrate that our approach maintains stable selection costs, reducing the token selection process by up to 110 ms, from 112.4 ms to just 2.5 ms at 512 tokens in Qwen2.5-VL. This structural efficiency successfully translates theoretical token reductions into tangible inference speedups while preserving highly competitive performance against state-of-the-art techniques under aggressive compression.

Jouwon Song, Woohyeong Kim, Kyeongbo Kong

LG Electronics, Seoul, Republic of Korea · Pusan National University, Busan, Republic of Korea

Efficient VLM Inference Large Vision-Language Models Visual Token Pruning Visual Token Compression

Jul 24, 2026cs.CV

LayoutLite: Token-Level Implicit Layout Analysis for Efficient Document OCR

End-to-end OCR systems based on vision-language models have achieved strong performance in complex document OCR, but their efficiency is limited by the large number of visual tokens produced from document images. Many of these tokens correspond to blank margins or visually redundant regions, yet directly applying generic visual token compression methods may remove OCR-critical fine-grained details. In this paper, we propose LayoutLite, a lightweight plug-and-play module for efficient document OCR. Instead of relying on explicit document layout detection, LayoutLite performs implicit layout analysis at the token level between the vision encoder and the language decoder. It aggregates multi-layer visual representations from the vision encoder, and predicts an importance score for each visual token with a lightweight scoring network. Low-information tokens are then removed before entering the language decoder while preserving the original spatial positional information of retained tokens. To train LayoutLite without human annotations, we cast token selection as a reinforcement learning problem and optimize it with a group-relative policy optimization objective driven by OCR output consistency, together with an auxiliary layout supervision signal to stabilize training. Experiments on OmniDocBench demonstrate that LayoutLite can substantially reduce visual token length and inference cost with negligible degradation in recognition quality. We further evaluate LayoutLite on two OCR-specialized VLMs, FireRed-OCR and Logics-Parsing-V2. Under up to 50% token compression, LayoutLite preserves almost the same score on both models while reducing prefill latency, FLOPs, and KV cache memory by over 40%, with only a small additional inference overhead. These results show that token-level implicit layout analysis is an effective and practical approach for accelerating VLM-based OCR systems.

Xudong Liu, Bicheng Wan, Yulin Jin

1Yuanli Technology (Beijing) Co., Ltd. · 2Beijing Normal University

Efficient VLM Inference Document Layout Analysis Optical Character Recognition Visual Token Pruning Visual Token Compression

Jul 23, 2026cs.CV

When Are Reasoning-Based Guardrails Not Efficient? ResponseGuard: A Fast Vision-Language Guard for Real-Time Moderation

A vision-language AI assistant returns its answer as a stream of generated tokens. Therefore, a safety guard that watches that answer has to keep up with the stream and stop a harmful reply before a user reads it. Recent vision-language guardrails instead generate a chain of thought before they issue a verdict. They believe that step-by-step reasoning yields a safer guard. This design makes the guard heavy and slow, since the model must decode many tokens for harmfulness detection. We pose the question of whether a vision-language guard really needs to reason in order to screen a response. We answer with a guard that has no chain. ResponseGuard reads a harmful verdict from a single pooled representation of the request, the response, and the image in one forward pass. Across a standard multimodal guardrail benchmark, our 2B ResponseGuard outperforms a recent 3B reasoning-based vision-language guard on response harmfulness detection, without any reasoning and at about 150 times lower time cost. On request harmfulness the reasoning guard retains an overall lead, and the remaining gap on both tracks sits on the image-only cells. We observe that the gap may stem from the frozen vision encoders that both designs use rather than from the missing chain. We have also found the reasoning guard directs almost none of its verdict attention to the image. Based on a single-pass detection, ResponseGuard can screen an answer sentence by sentence as it streams and stop a harmful answer before it finishes. For guarding the response of a vision-language model, a calibrated single-pass label may provide a sufficient safety signal. We fully release all source code, trained models, and datasets at https://github.com/ndb796/ResponseGuard.

Dongbin Na

POSTECH

VLM Evaluation Efficient VLM Inference Safety Filtering

Jul 22, 2026cs.CV

Look Less, Think Faster: Joint Token-Compute Adaptation for Multimodal LLMs

Multimodal Large Language Models (MLLMs) have recently demonstrated strong performance across vision-language tasks. However, their high inference cost, arising from both the large number of input visual tokens and the heavy computation of the large language model (LLM), remains a key barrier to practical deployment. Recent work attempts to reduce the cost by adaptively optimizing individual dimensions, e.g., pruning redundant visual tokens or skipping LLM layers and heads. Nonetheless, prior approaches typically treat these dimensions independently and overlook a fundamental coupling: the available compute resources must be dynamically allocated across all dimensions based on the input content. To bridge the gap, we propose SmartVL, a unified adaptive inference framework that jointly controls vision token number and model compute capability in response to varying input contents and compute budgets. SmartVL introduces a vision-side token controller that dynamically selects informative visual tokens and an LLM-side compute controller that adaptively adjusts LLM computation. Importantly, these controllers are trained to coordinate with each other so that the overall inference cost satisfies a target budget. To allow this joint scheduling, we connect the controllers using a shared budget encoding and leverage a differentiable latency estimator for end-to-end training. This design enables SmartVL to learn cross-stage allocation strategies that adapt to both input complexity and runtime compute constraints. Experiments across multiple MLLM benchmarks demonstrate that, with joint scheduling, SmartVL consistently outperforms prior adaptive methods and achieves superior accuracy-efficiency Pareto frontiers. Project page: https://www.schaterji.io/publications/2026/jointtokencompute.

Pengcheng Wang, Zhiquan Wang, Jayoung Lee +5

Purdue University, West Lafayette, IN, USA · University of Wisconsin–Madison, Madison, WI, USA · NVIDIA, Santa Clara, CA, USA

Efficient Multimodal Inference Efficient VLM Inference Multimodal Large Language Models Multimodal Foundation Models Visual Token Pruning Adaptive Computation

Jul 22, 2026cs.RO

SOPD-SocialNav: Selective On-Policy Distillation for Vision-Language Social Navigation

Vision-language models have shown strong potential for social robot navigation by leveraging rich semantic understanding of complex environments and human behaviors. However, large scale VLMs are difficult to deploy on resource-constrained robotic platforms, while lightweight VLMs often lack sufficient social reasoning capability. To address this problem, we propose SOPD-SocialNav, a selective on-policy distillation (SOPD) method that transfers social navigation knowledge from a large teacher VLM to a lightweight student VLM. SOPD introduces an entropy-based token selection mechanism that uses teacher uncertainty to identify socially informative decision tokens, while suppressing gradients from low-entropy tokens corresponding to trivial navigation states. A temperature-controlled Jensen-Shannon divergence objective is then used to align the student and teacher distributions on the selected tokens. Experiments on the SNEI and MUSON benchmarks demonstrate that SOPD consistently outperforms supervised fine-tuning, off-policy distillation, and standard on-policy distillation baselines in action prediction, perception consistency, and reasoning consistency. Real-world deployment on a Scout Mini robot further shows that the distilled model can generate more socially appropriate navigation behaviors in conversational and queuing scenarios. These results suggest that SOPD is an effective strategy for building lightweight yet socially aware VLM-based navigation systems.

Xinyu Zhang, Zishuo Wang, Ling Xiao

Graduate School of Information Science and Technology, Hokkaido University, Sapporo, Japan

Efficient VLM Inference VLM Distillation Social Robot Navigation On-Policy Distillation VLMs for Robotics

Jul 21, 2026eess.IV

BLUE: Semantics-Preserving Video Compression for Efficient Vision-Language Surveillance Analytics

Continuous surveillance video creates a growing storage, transmission, and inference burden for enterprise video analytics systems. While modern codecs such as H.265 reduce bitrate for human-viewable video, aggressive compression can degrade downstream computer-vision performance and does not necessarily reduce the number of vision-language model (VLM) inference calls required for semantic video understanding. This paper evaluates BLUE, a fixed-camera surveillance compression approach that suppresses static-background redundancy while preserving foreground activity, for its effect on VLM-based event and anomaly understanding. We compare raw H.265 and BLUE-compressed H.265 video on two surveillance datasets: VIRAT, comprising 227 paired event samples from 106 clips, and CHAD, comprising 54 human-activity anomaly clips. For each pair, the same frame index is evaluated using a VLM captioning pipeline, and outputs are scored against annotation-derived ground truth using a blind judging protocol. The results show no measurable degradation in semantic inference quality. On VIRAT, the mean VLM score remains effectively unchanged between raw H.265 and BLUE, with a mean difference of approximately -0.01 on a 0-10 scale. On CHAD, raw H.265 and BLUE obtain near-equivalent mean scores of 4.31 and 4.26, respectively. Compression saving is also uncorrelated with VLM score change on VIRAT (r = 0.004), indicating that higher BLUE compression does not predict semantic quality loss. Beyond storage reduction, BLUE increases the share of skip-heavy P-frames on CHAD from 1.4% to 53.2%, enabling an estimated 53% reduction in VLM calls through packet-size-based frame skipping. These findings suggest that BLUE functions as a machine-centric compression layer for surveillance video, reducing bandwidth and inference cost while preserving VLM semantic performance.

Shubham Baid, Akash James, Sahil Chachra +2

KGraph AI Solutions Pvt. Ltd., Bangalore, India – 560016

Efficient VLM Inference Video Compression Video Surveillance

Jul 21, 2026cs.CL

HPD-Parsing: Hierarchical Parallel Document Parsing

Efficient teamwork typically combines global coordination with parallel execution, a principle not yet fully reflected in unified Vision-Language Model (VLM)-based document parsers. Existing unified parsers process an entire page jointly but generate its output through a single token-by-token autoregressive trajectory, creating a sequential bottleneck that grows with document length. Such full-page sequential generation overlooks a key property of document parsing: layout must be analyzed globally, whereas block content can be parsed in parallel. Based on this observation, we introduce HPD-Parsing, which replaces full-page autoregressive generation with a Hierarchical Parallel Decoding paradigm. A main layout branch organizes the overall document structure and dynamically assigns block-level content decoding to concurrent branches, while progressive multi-token prediction (P-MTP) further reduces the decoding steps within each branch. Experiments on public benchmarks show that HPD-Parsing achieves 4,752 tokens per second, delivering

2.62\times

the throughput of the fastest existing document parsing model and

3.06\times

that of the vanilla autoregressive baseline, while maintaining competitive parsing accuracy. These results establish hierarchical parallel decoding as an effective alternative to full-page autoregressive generation, opening a new direction for efficient unified document parsing.

Shu Wei, Jingjing Wu, Lingshu Zhang +10

Efficient VLM Inference Document Parsing Parallel Decoding

Jul 21, 2026cs.CV

Bounding Boxes to Improve Small Language Model Performance on Vision-Based Grading Tasks

The deployment of Small Language Models (SLMs) in educational settings offers significant advantages in terms of privacy, cost, and scalability. However, SLMs often struggle with complex vision-based tasks, such as grading handwritten student exams, due to the high computational cost of processing large images and the visual distractions present on a full page. In this paper, we investigate whether cropping student responses using bounding boxes can improve the accuracy and computational efficiency of SLMs on a short-answer grading task. Using a dataset of scanned handwritten responses from the 2025 Australian Physics Olympiad, we evaluate the performance of several models ranging from 4B to 72B parameters under varying conditions of Chain of Thought (CoT) prompting and image cropping. Our results demonstrate that using bounding boxes significantly improves grading accuracy and reduces computational cost (FLOPs) across models. We conclude that bounding boxes are a crucial pre-processing step for deploying SLMs in large-scale, vision-based educational assessments.

Lachlan McGinness

Australian National University, Acton, ACT, Australia

VLM Evaluation Educational Assessment Efficient VLM Inference Small Language Models Automated Grading

Jul 20, 2026cs.CV

StepX-Edge: An On-Device UI Vision-Language Model via Architecture-Training-Deployment Co-Design

Deploying a vision-language model with full UI understanding on end devices has long been trapped between accuracy and efficiency: on one side is the accuracy bar for OCR, screen understanding, visual question answering, and element grounding; on the other is the strict compute, memory, and power budget of mobile chips. Existing work either trades one for the other, or stops at simulation without real-device validation. We present StepX-Edge, a 0.9B-parameter on-device UI vision-language model that resolves this tension through three-layer co-design of architecture, training, and deployment. Architecturally, UI-aware Layered Visual Encoding (ULVE) and a Progressive Dimensionality Projection (PDP) connector target the extreme aspect ratios and fine-grained perception of screens, while standard full attention throughout ensures native compatibility with mainstream mobile NPU operators. For training, the five-stage StepX-Curriculum framework is designed around our observation of mutual-promotion effects among UI subtasks, so that all four capabilities grow synergistically under a tight parameter budget rather than interfering. For deployment, a module-wise differentiated two-stage PTQ-to-QAT quantization scheme keeps the post-quantization accuracy loss within 1%. StepX-Edge achieves the strongest overall UI understanding among <=1B models, surpassing all 2B-2.3B baselines on ScreenQA (88.76 F1) and Chinese OCRBench v2 (57.25), and matching 1.3B-2.3B general VLMs on RefCOCO (92.0%) and OCRBench v1 (831) with far fewer parameters. After W4A16+KV8 quantization, the model runs stably on Snapdragon 8 Gen5 devices with ~0.84 s TTFT, 98 tok/s decode, and 1.4 GB peak memory. We will open-source the training data, the full training recipe, and the quantization deployment pipeline.

Yin Wang, Haotian Hu, Jineng Han +4

StepX Team

Vision-Language Models VLM Quantization Efficient VLM Inference Curriculum Learning On-Device Inference

Jul 19, 2026cs.CV

Searching for Task-Specific Vision Paths: Evolutionary Block Pruning Across Vision-Language Models

Vision-language models normally execute the same complete vision encoder for every question, even when OCR, counting, object, attribute, and spatial queries may not require identical computation. We study whether fixed-budget combinations of vision blocks can be skipped without fine-tuning. A shared K-block route skips one searched set of exactly K blocks for every question, while a capability-specific K-block policy selects one same-size route using a known capability label. We introduce a source-balanced evolutionary search and compare it with independent ranking, contiguous removal, and random routes at matched budgets. Experiments use Qwen2.5-VL-3B-Instruct, SmolVLM2-2.2B-Instruct, and an 876-example image-disjoint selection split. Search transfers across architectures: on SmolVLM2, the searched shared four-block route beats independent construction by 4.91 percentage points. Capability specialization is less stable. On Qwen, the six-block capability policy beats the shared route by 2.17 points, driven by a 7.10-point OCR gain. On sealed IIIT5K, however, the SmolVLM2 OCR-specific route trails its shared route by 13.6 points. Combinatorial search reliably improves route construction, but capability labels do not define universally transferable vision pathways.

Tarun Tomar

Efficient VLM Inference Neural Network Pruning

Jul 17, 2026cs.CV

Efficient Frame Selection for Long Videos at Test Time with Attention-Based MLLM Selectors

Understanding long videos with multimodal large language models (MLLMs) requires selecting a compact set of frames from thousands of candidates, yet identifying the right frames seemingly requires understanding the video first. We resolve this circular dependency with a simple observation: cross-modal attention at validation-selected extraction layers in MLLMs already provides query-relevant frame evidence without requiring autoregressive generation. We exploit this property to build DAFS (Dynamic Attention-based Budget-aware Frame Selection), a training-free frame selector. A lightweight MLLM selector, even with only 2B parameters, can extract frame-level evidence by converting selected-layer attention into relevance scores through query-conditioned aggregation. This enables cross-frame comparison without autoregressive decoding. To handle the selector's own context constraint, we formulate the joint allocation of candidate pool size and per-frame token budget as a discrete optimization problem solved by dynamic programming. Under a 32-frame budget, our selector improves over uniform sampling by up to 6.4 points on Video-MME and outperforms prior training-based selectors under matched frame budgets, while generalizing across selector and answerer backbones, and across tasks, without retraining.

Yilin Wang, Xiangxi Zheng, Dongxing Mao +6

ZJU · NJU · CSU +2

Efficient VLM Inference Multimodal Large Language Models Video Frame Selection Query-Aware Video Frame Selection

Jul 16, 2026cs.CV

VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding

Recent advances in video understanding have spanned motion, long video, and streaming interaction, driving this field toward real-world applications. Despite this progress, current open-source models remain limited in several ways. They often struggle to generalize across diverse video types, making them effective only in specific domains. High computational demands further restrict their efficiency and scalability. Moreover, most models are only partially open, with key components such as training code, strategy, or datasets unavailable, which hinders reproducibility and slows community-driven development. To address these issues, we introduce VideoChat3, a fully open, efficient, and generalist video-centric MLLM. VideoChat3 advances video understanding through two complementary designs. For efficiency, we introduce Inflated 3D Vision Transformer (I3D-ViT) and Adaptive Frame Resolution for Streaming Video Perception, which enables efficient spatiotemporal representation and reduces the cost of processing video inputs during training and inference. For effectiveness, we develop a scalable video data synthesis pipeline that curates three diverse, high-quality training datasets: VideoChat3-Academic2M, VideoChat3-LV116K, and VideoChat3-OL617K, covering general, long-form, and streaming video scenarios, improving the model's generalization across domains. By integrating these designs, VideoChat3 achieves a rare balance of broad generalization and computational efficiency. Experiments across general, long-form, and streaming benchmarks demonstrate that VideoChat3 surpasses prior open-source models with equal or larger parameter counts with only 4B parameters and higher efficiency.

Xinhao Li, Yuhan Zhu, Xiangyu Zeng +24

Nanjing University · Nanyang Technological University · Shanghai AI Laboratory +1

Video Understanding Streaming Video Understanding Efficient VLM Inference Video-Language Models