DiD It in 87 Minutes: A Label-Free Softmax-to-Linear Adaptation of Vision Transformers for Object Detection

• URL: http://arxiv.org/abs/2608.22368v1
• Date: Sun, 23 Aug 2026 11:21:16 GMT
• Title: DiD It in 87 Minutes: A Label-Free Softmax-to-Linear Adaptation of Vision Transformers for Object Detection
• Abstract summary: Detector-Interface Distillation (DiD) is a label-free conversion method that trains the linear-attention backbone by aligning interface tensors with those of a frozen Softmax teacher.<n>On DOTA-v1.5, DiD substantially outperforms established baselines and matches supervised, fully trained linear models.
• Score: 27.227554135787543
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While linear attention is a compelling mechanism for high-resolution object detection due to its reduced cost for global token mixing, converting the Softmax-attention ViT backbone of a trained detector into a linear-attention one is not a trivial drop-in replacement. Directly swapping the attention operator leads to severe performance degradation, and generic label-free distillation, though effective for classification, often fails on detection tasks. We argue that the central challenge is \textit{detector-interface preservation}: the converted backbone must reproduce the exact feature tensors expected by the fixed downstream detector, rather than merely imitating internal Softmax hidden states. To address this, we introduce Detector-Interface Distillation (DiD), a label-free conversion method that exclusively trains the linear-attention backbone by aligning detector-facing interface tensors with those of a frozen Softmax teacher. On DOTA-v1.5, DiD substantially outperforms established baselines and matches supervised, fully trained linear models. Adaptation completes in roughly 87 minutes on 4 GPUs, and the linearized backbone cuts inference latency by ~62% and peak memory by ~49%. We hope our findings offer the community a simple, label-free route to reusing trained Softmax detectors as efficient linear ones, and encourage interface-aware objectives in future architecture-conversion work.

Related papers

• Visual Geometry Transformer in the Wild: Distractor-Free 3D Reconstruction [77.75911117889915]
  Current end-to-end multi-view 3D reconstruction methods achieve impressive results, but rely on a restrictive static assumption.<n>This reliance on idealized inputs causes even the most advanced methods to fail in real-world settings.<n>We propose Visual Geometry Transformer in the Wild, an end-to-end framework for robust reconstruction from inconsistent views.
  arXiv  Detail & Related papers  (2026-06-22T02:52:22Z)
• Need We Teach Foundation Models What is a Generative Image? Gradient-Free Generative Artifact Detection via Analytic Spectral Adaptation [3.881211374324378]
  Adapting foundation models to detect generative artifacts via gradient-based updates compromises their intrinsic representations.<n>We propose a gradient-free methodology reframing detection from binary classification to an out-of-distribution (OOD) anomaly measurement problem.
  arXiv  Detail & Related papers  (2026-06-03T13:04:23Z)
• Amped: Adaptive Multi-stage Non-edge Pruning for Edge Detection [53.59823657644554]
  We propose an Adaptive Multi-stage non-edge Pruning framework for Edge Detection(Amped)<n>Amped identifies high-confidence non-edge tokens and removes them as early as possible to substantially reduce computation.<n>To mitigate the structural complexity of existing edge detection networks, we introduce a simple yet high-performance Transformer-based model, termed Streamline Edge Detector(SED)
  arXiv  Detail & Related papers  (2026-03-29T12:26:35Z)
• Axle Sensor Fusion for Online Continual Wheel Fault Detection in Wayside Railway Monitoring [1.1199585259018456]
  This work presents a semantic-aware, label-efficient continual learning framework for railway fault diagnostics.<n>The model detects minor imperfections due to flats and polygonization, while adapting to evolving operational conditions.
  arXiv  Detail & Related papers  (2026-02-18T00:14:18Z)
• TinyDef-DETR: A Transformer-Based Framework for Defect Detection in Transmission Lines from UAV Imagery [12.48571944931548]
  TinyDef-DETR is a framework designed to achieve accurate and efficient detection of transmission line defects from UAV-acquired images.<n>The model integrates four major components: an edge-enhanced ResNet backbone to strengthen boundary-sensitive representations, a stride-free space-to-depth module to enable detail-preserving downsampling, and a Focaler-Wise-SIoU regression loss to improve the localization of small and difficult objects.
  arXiv  Detail & Related papers  (2025-09-07T12:36:33Z)
• CARE Transformer: Mobile-Friendly Linear Visual Transformer via Decoupled Dual Interaction [77.8576094863446]
  We propose a new detextbfCoupled dutextbfAl-interactive lineatextbfR atttextbfEntion (CARE) mechanism. We first propose an asymmetrical feature decoupling strategy that asymmetrically decouples the learning process for local inductive bias and long-range dependencies. By adopting a decoupled learning way and fully exploiting complementarity across features, our method can achieve both high efficiency and accuracy.
  arXiv  Detail & Related papers  (2024-11-25T07:56:13Z)
• OVLW-DETR: Open-Vocabulary Light-Weighted Detection Transformer [63.141027246418]
  We propose Open-Vocabulary Light-Weighted Detection Transformer (OVLW-DETR), a deployment friendly open-vocabulary detector with strong performance and low latency. We provide an end-to-end training recipe that transferring knowledge from vision-language model (VLM) to object detector with simple alignment. Experimental results demonstrate that the proposed approach is superior over existing real-time open-vocabulary detectors on standard Zero-Shot LVIS benchmark.
  arXiv  Detail & Related papers  (2024-07-15T12:15:27Z)
• MutDet: Mutually Optimizing Pre-training for Remote Sensing Object Detection [36.478530086163744]
  We propose a novel Mutually optimizing pre-training framework for remote sensing object Detection, dubbed as MutDet. MutDet fuses the object embeddings and detector features bidirectionally in the last encoder layer, enhancing their information interaction. Experiments on various settings show new state-of-the-art transfer performance.
  arXiv  Detail & Related papers  (2024-07-13T15:28:15Z)
• DTAAD: Dual Tcn-Attention Networks for Anomaly Detection in Multivariate Time Series Data [0.0]
  We propose an anomaly detection and diagnosis model, DTAAD, based on Transformer and Dual Temporal Convolutional Network (TCN) scaling methods and feedback mechanisms are introduced to improve prediction accuracy and expand correlation differences. Our experiments on seven public datasets validate that DTAAD exceeds the majority of currently advanced baseline methods in both detection and diagnostic performance.
  arXiv  Detail & Related papers  (2023-02-17T06:59:45Z)
• Self-Supervised Training with Autoencoders for Visual Anomaly Detection [61.62861063776813]
  We focus on a specific use case in anomaly detection where the distribution of normal samples is supported by a lower-dimensional manifold. We adapt a self-supervised learning regime that exploits discriminative information during training but focuses on the submanifold of normal examples. We achieve a new state-of-the-art result on the MVTec AD dataset -- a challenging benchmark for visual anomaly detection in the manufacturing domain.
  arXiv  Detail & Related papers  (2022-06-23T14:16:30Z)
• Efficient Decoder-free Object Detection with Transformers [75.00499377197475]
  Vision transformers (ViTs) are changing the landscape of object detection approaches. We propose a decoder-free fully transformer-based (DFFT) object detector. DFFT_SMALL achieves high efficiency in both training and inference stages.
  arXiv  Detail & Related papers  (2022-06-14T13:22:19Z)
• Integral Migrating Pre-trained Transformer Encoder-decoders for Visual Object Detection [78.2325219839805]
  imTED improves the state-of-the-art of few-shot object detection by up to 7.6% AP. Experiments on MS COCO dataset demonstrate that imTED consistently outperforms its counterparts by 2.8%.
  arXiv  Detail & Related papers  (2022-05-19T15:11:20Z)

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.