JetViT: Efficient High-Resolution Vision Transformer with Post-Training Attention Search

• URL: http://arxiv.org/abs/2605.26636v1
• Date: Tue, 26 May 2026 07:17:27 GMT
• Title: JetViT: Efficient High-Resolution Vision Transformer with Post-Training Attention Search
• Abstract summary: JetViT is a novel family of hybrid-architecture Vision Transformer (ViT) models.<n>On the NVIDIA H100 GPU, JetViT achieves up to 1.79x higher throughput and up to 44.81% lower latency without sacrificing accuracy.
• Score: 49.62748530356587
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We introduce JetViT, a novel family of hybrid-architecture Vision Transformer (ViT) models that match the accuracy of state-of-the-art full-attention vision foundation models while achieving substantially higher inference efficiency on high-resolution images. At the core of our approach is Post-Training Attention Search, a post-training acceleration framework that converts pre-trained full-attention ViTs into efficient hybrid-attention variants by identifying and replacing redundant full-attention blocks with linear or window-attention blocks. By inheriting the MLP and attention weights from the base model, Post-Training Attention Search efficiently explores the architectural design space through three key steps: (1) optimizing the linear-attention block design; (2) finding the best combination of linear-attention and window-attention blocks; and (3) identifying and preserving critical full-attention blocks. We evaluate JetViT on two representative high-resolution vision foundation models, DINOv3 and DepthAnythingV2. On the NVIDIA H100 GPU, JetViT achieves up to 1.79x higher throughput and up to 44.81% lower latency without sacrificing accuracy. We will release our code and accelerated ViT models soon.

Related papers

• MicroViTv2: Beyond the FLOPS for Edge Energy-Friendly Vision Transformers [6.427162946484909]
  Vision Transformer (ViT) achieves remarkable accuracy across visual tasks but remains computationally expensive for edge deployment.<n>This paper presents MicroViTv2, a lightweight Vision Transformer optimized for real-device efficiency.
  arXiv  Detail & Related papers  (2026-05-11T07:54:27Z)
• HY-Embodied-0.5: Embodied Foundation Models for Real-World Agents [71.09493646932046]
  We introduce HY-Embodied-0.5, a family of foundation models specifically designed for real-world embodied agents.<n>The HY-Embodied-0.5 suite comprises two primary variants: an efficient model with 2B activated parameters designed for edge deployment, and a powerful model with 32B activated parameters targeted for complex reasoning.<n>Our MoT-2B model outperforms similarly sized state-of-the-art models on 16 benchmarks, while the 32B variant achieves performance comparable to frontier models such as Gemini 3.0 Pro.
  arXiv  Detail & Related papers  (2026-04-08T17:59:48Z)
• FastVGGT: Training-Free Acceleration of Visual Geometry Transformer [83.67766078575782]
  VGGT is a state-of-the-art feed-forward visual geometry model.<n>We propose FastVGGT, which leverages token merging in the 3D domain through a training-free mechanism for accelerating VGGT.<n>With 1000 input images, FastVGGT achieves a 4x speedup over VGGT while mitigating error accumulation in long-sequence scenarios.
  arXiv  Detail & Related papers  (2025-09-02T17:54:21Z)
• VMoBA: Mixture-of-Block Attention for Video Diffusion Models [29.183614108287276]
  This paper introduces Video Mixture of Block Attention (VMoBA), a novel attention mechanism specifically adapted for Video Diffusion Models (VDMs)<n>Motivated by an in-depth analysis of attention patterns within pre-trained video transformers, VMoBA enhances the original MoBA framework with three key modifications.<n>Experiments demonstrate that VMoBA significantly accelerates the training of VDMs on longer sequences, achieving 2.92x FLOPs and 1.48x latency speedup.
  arXiv  Detail & Related papers  (2025-06-30T13:52:31Z)
• ViT-Linearizer: Distilling Quadratic Knowledge into Linear-Time Vision Models [39.520825264698374]
  Vision Transformers (ViTs) have delivered remarkable progress through global self-attention, yet their quadratic complexity can become prohibitive for high-resolution inputs.<n>We present ViT-Linearizer, a cross-architecture distillation framework that transfers rich ViT representations into a linear-time, recurrent-style model.<n>Our results underscore the good potential of RNN-based solutions for large-scale visual tasks, bridging the gap between theoretical efficiency and real-world practice.
  arXiv  Detail & Related papers  (2025-03-30T15:35:24Z)
• Exploring Lightweight Hierarchical Vision Transformers for Efficient Visual Tracking [69.89887818921825]
  HiT is a new family of efficient tracking models that can run at high speed on different devices. HiT achieves 64.6% AUC on the LaSOT benchmark, surpassing all previous efficient trackers.
  arXiv  Detail & Related papers  (2023-08-14T02:51:34Z)
• GOHSP: A Unified Framework of Graph and Optimization-based Heterogeneous Structured Pruning for Vision Transformer [76.2625311630021]
  Vision transformers (ViTs) have shown very impressive empirical performance in various computer vision tasks. To mitigate this challenging problem, structured pruning is a promising solution to compress model size and enable practical efficiency. We propose GOHSP, a unified framework of Graph and Optimization-based Structured Pruning for ViT models.
  arXiv  Detail & Related papers  (2023-01-13T00:40:24Z)
• ViTALiTy: Unifying Low-rank and Sparse Approximation for Vision Transformer Acceleration with a Linear Taylor Attention [23.874485033096917]
  Vision Transformer (ViT) has emerged as a competitive alternative to convolutional neural networks for various computer vision applications. We propose a first-of-its-kind algorithm- hardware codesigned framework, dubbed ViTALiTy, for boosting the inference efficiency of ViTs. ViTALiTy unifies both low-rank and sparse components of the attention in ViTs.
  arXiv  Detail & Related papers  (2022-11-09T18:58:21Z)
• EfficientViT: Multi-Scale Linear Attention for High-Resolution Dense Prediction [67.11722682878722]
  This work presents EfficientViT, a new family of high-resolution vision models with novel multi-scale linear attention. Our multi-scale linear attention achieves the global receptive field and multi-scale learning. EfficientViT delivers remarkable performance gains over previous state-of-the-art models.
  arXiv  Detail & Related papers  (2022-05-29T20:07:23Z)
• Twins: Revisiting Spatial Attention Design in Vision Transformers [81.02454258677714]
  In this work, we demonstrate that a carefully-devised yet simple spatial attention mechanism performs favourably against the state-of-the-art schemes. We propose two vision transformer architectures, namely, Twins-PCPVT and Twins-SVT. Our proposed architectures are highly-efficient and easy to implement, only involving matrix multiplications that are highly optimized in modern deep learning frameworks.
  arXiv  Detail & Related papers  (2021-04-28T15:42:31Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.