English
Related papers

Related papers: ButterflyViT: 354$\times$ Expert Compression for E…

200 papers

Quantization has significantly improved the compute and memory efficiency of Large Language Model (LLM) training. However, existing approaches still rely on accumulating their updates in high-precision: concretely, gradient updates must be…

Computation and Language · Computer Science 2026-01-30 Mahdi Nikdan , Amir Zandieh , Dan Alistarh , Vahab Mirrokni

The demand for edge AI in vision-language tasks requires models that achieve real-time performance on resource-constrained devices with limited power and memory. This paper proposes two adaptive compression techniques -- Sparse Temporal…

Computer Vision and Pattern Recognition · Computer Science 2025-11-25 Md Tasnin Tanvir , Soumitra Das , Sk Md Abidar Rahaman , Ali Shiri Sichani

Recent advances in vision-language models (VLMs) have demonstrated the advantages of processing images at higher resolutions and utilizing multi-crop features to preserve native resolution details. However, despite these improvements,…

Computer Vision and Pattern Recognition · Computer Science 2024-10-16 Rahul Thapa , Kezhen Chen , Ian Covert , Rahul Chalamala , Ben Athiwaratkun , Shuaiwen Leon Song , James Zou

Vision Transformers (ViTs) have delivered remarkable progress through global self-attention, yet their quadratic complexity can become prohibitive for high-resolution inputs. In this work, we present ViT-Linearizer, a cross-architecture…

Computer Vision and Pattern Recognition · Computer Science 2026-02-27 Guoyizhe Wei , Rama Chellappa

Token compression aims to speed up large-scale vision transformers (e.g. ViTs) by pruning (dropping) or merging tokens. It is an important but challenging task. Although recent advanced approaches achieved great success, they need to…

Computer Vision and Pattern Recognition · Computer Science 2023-05-30 Mengzhao Chen , Wenqi Shao , Peng Xu , Mingbao Lin , Kaipeng Zhang , Fei Chao , Rongrong Ji , Yu Qiao , Ping Luo

Vision Transformer models process input images by dividing them into a spatially regular grid of equal-size patches. Conversely, Transformers were originally introduced over natural language sequences, where each token represents a subword…

Computer Vision and Pattern Recognition · Computer Science 2023-04-28 Tomer Ronen , Omer Levy , Avram Golbert

Steered-Mixtures-of-Experts (SMoE) models provide sparse, edge-aware representations, applicable to many use-cases in image processing. This includes denoising, super-resolution and compression of 2D- and higher dimensional pixel data.…

Image and Video Processing · Electrical Eng. & Systems 2022-07-26 Elvira Fleig , Jonas Geistert , Erik Bochinski , Rolf Jongebloed , Thomas Sikora

Recent works have demonstrated that transformer can achieve promising performance in computer vision, by exploiting the relationship among image patches with self-attention. While they only consider the attention in a single feature layer,…

Computer Vision and Pattern Recognition · Computer Science 2023-06-12 Nannan Li , Yaran Chen , Weifan Li , Zixiang Ding , Dongbin Zhao

Structural re-parameterization is a general training scheme for Convolutional Neural Networks (CNNs), which achieves performance improvement without increasing inference cost. As Vision Transformers (ViTs) are gradually surpassing CNNs in…

Computer Vision and Pattern Recognition · Computer Science 2023-08-28 Yongqi Huang , Peng Ye , Xiaoshui Huang , Sheng Li , Tao Chen , Tong He , Wanli Ouyang

Recent advances in multimodal foundation models unifying image understanding and generation have opened exciting avenues for tackling a wide range of vision-language tasks within a single framework. Despite progress, existing unified models…

Computer Vision and Pattern Recognition · Computer Science 2025-06-10 Ying Shen , Zhiyang Xu , Jiuhai Chen , Shizhe Diao , Jiaxin Zhang , Yuguang Yao , Joy Rimchala , Ismini Lourentzou , Lifu Huang

Mixture-of-Experts(MoE) Vision-Language Models (VLMs) offer remarkable performance but incur prohibitive memory and computational costs, making compression essential. Post-Training Quantization (PTQ) is an effective training-free technique…

Computer Vision and Pattern Recognition · Computer Science 2026-02-03 Guangshuo Qin , Zhiteng Li , Zheng Chen , Weihang Zhang , Linghe Kong , Yulun Zhang

Point cloud compression has garnered significant interest in computer vision. However, existing algorithms primarily cater to human vision, while most point cloud data is utilized for machine vision tasks. To address this, we propose a…

Computer Vision and Pattern Recognition · Computer Science 2024-06-04 Lei Liu , Zhihao Hu , Zhenghao Chen

Large multimodal Mixture-of-Experts (MoEs) effectively scale the model size to boost performance while maintaining fixed active parameters. However, previous works primarily utilized full-precision experts during sparse up-cycling. Despite…

Computer Vision and Pattern Recognition · Computer Science 2026-01-08 Hongyu Wang , Jiayu Xu , Ruiping Wang , Yan Feng , Yitao Zhai , Peng Pei , Xunliang Cai , Xilin Chen

Ternary Vision Transformers offer substantial model compression, however state-of-the-art methods only ternarize the encoder layers, leaving patch embeddings, LayerNorm parameters, and classifier heads in full precision. In compact models…

Computer Vision and Pattern Recognition · Computer Science 2026-05-21 Szymon Ruciński , Pietro Bonazzi , Engin Türetken , Simon Narduzzi , Michele Magno , Nadim Maamari

Transformer-based architectures have demonstrated remarkable success across various domains, but their deployment on edge devices remains challenging due to high memory and computational demands. In this paper, we introduce a novel Reuse…

Computer Vision and Pattern Recognition · Computer Science 2024-12-04 Seul-Ki Yeom , Tae-Ho Kim

Post-Training Quantization (PTQ) has emerged as an effective technique for alleviating the substantial computational and memory overheads of Vision-Language Models (VLMs) by compressing both weights and activations without retraining the…

Computer Vision and Pattern Recognition · Computer Science 2026-03-02 Chenwei Jia , Baoting Li , Xuchong Zhang , Mingzhuo Wei , Bochen Lin , Hongbin Sun

The QMoE model provides a practical approach for compression of massive Mixture-of-Experts (MoE) models. QMoE offers a solution geared towards memory limitations that often reach terabyte scales, and it has the advantage of working with…

Performance · Computer Science 2025-09-30 Jack Cashman , Jiaqi Nie

The number of mobile robots with constrained computing resources that need to execute complex machine learning models has been increasing during the past decade. Commonly, these robots rely on edge infrastructure accessible over wireless…

Machine Learning · Computer Science 2022-07-22 Timotheos Souroulla , Alberto Hata , Ahmad Terra , Özer Özkahraman , Rafia Inam

The accelerated MRI reconstruction process presents a challenging ill-posed inverse problem due to the extensive under-sampling in k-space. Recently, Vision Transformers (ViTs) have become the mainstream for this task, demonstrating…

Image and Video Processing · Electrical Eng. & Systems 2024-12-17 Yucong Meng , Zhiwei Yang , Yonghong Shi , Zhijian Song

Federated learning on edge devices must cope with non-IID client data and tight memory budgets. Adaptive optimizers like Adam stabilize training under data heterogeneity but require storing full-precision momentum and variance states, often…

Machine Learning · Computer Science 2026-05-19 Vedant Waykole , Haroon R. Lone