Wang, HaoqiZhang, TongSalzmann, MathieuLeonardis, ARicci, ERoth, SRussakovsky, OSattler, TVarol, G2025-01-312025-01-312025-01-312025-01-0110.1007/978-3-031-72667-5_2https://infoscience.epfl.ch/handle/20.500.14299/246140WOS:001346380800002Vision Transformer models trained on large-scale datasets, although effective, often exhibit artifacts in the patch token they extract. While such defects can be alleviated by re-training the entire model with additional classification tokens, the underlying reasons for the presence of these tokens remain unclear. In this paper, we conduct a thorough investigation of this phenomenon, combining theoretical analysis with empirical observations. Our findings reveal that these artifacts originate from the pre-trained network itself, specifically stemming from the leading left singular vector of the network's weights. Furthermore, to mitigate these defects, we propose a novel fine-tuning smooth regularization that rectifies structural deficiencies using only a small dataset, thereby avoiding the need for complete re-training. We validate our method on various downstream tasks, including unsupervised segmentation, classification, supervised segmentation, and depth estimation, demonstrating its effectiveness in improving model performance. Codes and checkpoints are available at https://github.com/haoqiwang/sinder.EnglishDINOv2Singular DefectUnsupervised Segmentationtext::conference output::conference proceedings::conference paper