Papers › Vision Transformers with Patch Diversification

Vision Transformers with Patch Diversification

26 Apr 2021arXiv:2104.12753archive 2025-07-28

Chengyue Gong, Dilin Wang, Meng Li, Vikas Chandra, Qiang Liu

Vision transformer has demonstrated promising performance on challenging computer vision tasks. However, directly training the vision transformers may yield unstable and sub-optimal results. Recent works propose to improve the performance of the vision transformers by modifying the transformer structures, e.g., incorporating convolution layers. In contrast, we investigate an orthogonal approach to stabilize the vision transformer training without modifying the networks. We observe the instability of the training can be attributed to the significant similarity across the extracted patch representations. More specifically, for deep vision transformers, the self-attention blocks tend to map different patches into similar latent representations, yielding information loss and performance degradation. To alleviate this problem, in this work, we introduce novel loss functions in vision transformer training to explicitly encourage diversity across patch representations for more discriminative feature extraction. We empirically show that our proposed techniques stabilize the training and allow us to train wider and deeper vision transformers. We further show the diversified features significantly benefit the downstream tasks in transfer learning. For semantic segmentation, we enhance the state-of-the-art (SOTA) results on Cityscapes and ADE20k. Our code is available at https://github.com/ChengyueGongR/PatchVisionTransformer.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

ChengyueGongR/PatchVisionTransformer officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

DiversityImage ClassificationSemantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Semantic Segmentation ADE20K PatchDiverse + Swin-L (multi-scale test, upernet, ImageNet22k pretrain) Validation mIoU 54.4 #61 of 235 Archive leaderboard report
Semantic Segmentation ADE20K val PatchDiverse + Swin-L (multi-scale test, upernet, ImageNet22k pretrain) mIoU 54.4% #32 of 95 Archive leaderboard report
Semantic Segmentation Cityscapes val PatchDiverse + Swin-L (multi-scale test, upernet, ImageNet22k pretrain) mIoU 83.6% #24 of 99 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AttentionConvolutionDense ConnectionsLayer NormalizationLinear LayerMulti-Head AttentionResidual ConnectionSoftmaxVision Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections