LEAP: Layer-skipping Efficiency via Adaptive Progression for Vision Transformer Distillation

• URL: http://arxiv.org/abs/2606.19483v1
• Date: Wed, 17 Jun 2026 18:14:51 GMT
• Title: LEAP: Layer-skipping Efficiency via Adaptive Progression for Vision Transformer Distillation
• Abstract summary: We propose LEAP: Layer-skipping Efficiency via Adaptive Progression, a training curriculum for ViT feature-based knowledge distillation.<n>Our results demonstrate that this paradigm significantly accelerates convergence through adaptive difficulty selection.<n> curriculum enables 25.1% savings in training FLOPs and 21% savings in training time on ImageNet-100.
• Score: 21.96454507185653
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision Foundation Models (VFMs) with Vision Transformer (ViT) backbones, such as DINOv2, have become essential for downstream tasks like object recognition and semantic segmentation. The immense computational requirements of backbones often necessitate distillation into smaller architectures for edge deployment. Feature-based knowledge distillation (KD) often suffers from the teacher-student gap; the student struggles to imitate teacher's complex feature map due to its limited capacity. To mitigate this bottleneck, we propose LEAP: Layer-skipping Efficiency via Adaptive Progression, a training curriculum for ViT feature-based knowledge distillation. By utilizing the teacher's intermediate feature maps as a sequence of progressively more difficult targets, our curriculum allows the student to build a foundational representation before tackling higher-level abstractions. Our results demonstrate that this paradigm significantly accelerates convergence through adaptive difficulty selection across various student model sizes and dataset scales. With our curriculum, the LEAP-distilled ViT-S achieves 90.1% accuracy on ImageNet-100, a +12.24% improvement compared with baseline. On ImageNet-1K, LEAP achieves +3.84% and +7.75% improvement for the instance retrieval task on the Oxford and Paris datasets, respectively. Furthermore, the curriculum enables 25.1% savings in training FLOPs and 21% savings in training time on ImageNet-100 by implementing early-stopping for teacher inference during the initial stages of training. Code is available at https://github.com/KevinZ0217/LEAP

Related papers

• Next-Embedding Prediction Makes Strong Vision Learners [68.55755328850634]
  We train models to generate embeddings to perform predictive tasks directly.<n>Next-Embedding Predictive Autoregression (NEPA) achieves strong results across tasks.<n>We believe generative pretraining from embeddings provides a simple, scalable, and potentially modality-agnostic alternative to visual self-supervised learning.
  arXiv  Detail & Related papers  (2025-12-18T18:59:58Z)
• Underlying Semantic Diffusion for Effective and Efficient In-Context Learning [113.4003355229632]
  Underlying Semantic Diffusion (US-Diffusion) is an enhanced diffusion model that boosts underlying semantics learning, computational efficiency, and in-context learning capabilities.<n>We present a Feedback-Aided Learning (FAL) framework, which leverages feedback signals to guide the model in capturing semantic details.<n>We also propose a plug-and-play Efficient Sampling Strategy (ESS) for dense sampling at time steps with high-noise levels.
  arXiv  Detail & Related papers  (2025-03-06T03:06:22Z)
• ScaleKD: Strong Vision Transformers Could Be Excellent Teachers [15.446480934024652]
  We present a simple and effective knowledge distillation method, called ScaleKD. Our method can train student backbones that span across a variety of convolutional neural network (CNN), multi-layer perceptron (MLP), and ViT architectures on image classification datasets. When scaling up the size of teacher models or their pre-training datasets, our method showcases the desired scalable properties.
  arXiv  Detail & Related papers  (2024-11-11T08:25:21Z)
• Augmentation-Free Dense Contrastive Knowledge Distillation for Efficient Semantic Segmentation [16.957139277317005]
  Augmentation-free Dense Contrastive Knowledge Distillation (Af-DCD) is a new contrastive distillation learning paradigm. Af-DCD trains compact and accurate deep neural networks for semantic segmentation applications.
  arXiv  Detail & Related papers  (2023-12-07T09:37:28Z)
• Local Masking Meets Progressive Freezing: Crafting Efficient Vision Transformers for Self-Supervised Learning [0.0]
  We present an innovative approach to self-supervised learning for Vision Transformers (ViTs) This method focuses on enhancing the efficiency and speed of initial layer training in ViTs. Our approach employs a novel multi-scale reconstruction process that fosters efficient learning in initial layers.
  arXiv  Detail & Related papers  (2023-12-02T11:10:09Z)
• Effective Whole-body Pose Estimation with Two-stages Distillation [52.92064408970796]
  Whole-body pose estimation localizes the human body, hand, face, and foot keypoints in an image. We present a two-stage pose textbfDistillation for textbfWhole-body textbfPose estimators, named textbfDWPose, to improve their effectiveness and efficiency.
  arXiv  Detail & Related papers  (2023-07-29T03:49:28Z)
• Getting More Juice Out of Your Data: Hard Pair Refinement Enhances Visual-Language Models Without Extra Data [122.282521548393]
  Contrastive Language-Image Pre-training (CLIP) has become the standard for cross-modal image-text representation learning.<n>We introduce HELIP, a cost-effective strategy that improves CLIP models by exploiting challenging text-image pairs within existing datasets in continuous training.
  arXiv  Detail & Related papers  (2023-05-09T07:00:17Z)
• DisWOT: Student Architecture Search for Distillation WithOut Training [0.0]
  We explore a novel training-free framework to search for the best student architectures for a given teacher. Our work first empirically show that the optimal model under vanilla training cannot be the winner in distillation. Our experiments on CIFAR, ImageNet and NAS-Bench-201 demonstrate that our technique achieves state-of-the-art results on different search spaces.
  arXiv  Detail & Related papers  (2023-03-28T01:58:45Z)
• A Simple and Generic Framework for Feature Distillation via Channel-wise Transformation [35.233203757760066]
  We propose a learnable nonlinear channel-wise transformation to align the features of the student and the teacher model. Our method achieves significant performance improvements in various computer vision tasks.
  arXiv  Detail & Related papers  (2023-03-23T12:13:29Z)
• EfficientTrain: Exploring Generalized Curriculum Learning for Training Visual Backbones [80.662250618795]
  This paper presents a new curriculum learning approach for the efficient training of visual backbones (e.g., vision Transformers) As an off-the-shelf method, it reduces the wall-time training cost of a wide variety of popular models by >1.5x on ImageNet-1K/22K without sacrificing accuracy.
  arXiv  Detail & Related papers  (2022-11-17T17:38:55Z)
• Offline Visual Representation Learning for Embodied Navigation [50.442660137987275]
  offline pretraining of visual representations with self-supervised learning (SSL) Online finetuning of visuomotor representations on specific tasks with image augmentations under long learning schedules.
  arXiv  Detail & Related papers  (2022-04-27T23:22:43Z)
• Self-Supervised Pre-Training for Transformer-Based Person Re-Identification [54.55281692768765]
  Transformer-based supervised pre-training achieves great performance in person re-identification (ReID) Due to the domain gap between ImageNet and ReID datasets, it usually needs a larger pre-training dataset to boost the performance. This work aims to mitigate the gap between the pre-training and ReID datasets from the perspective of data and model structure.
  arXiv  Detail & Related papers  (2021-11-23T18:59:08Z)

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.