Papers › Generalized Parametric Contrastive Learning

Generalized Parametric Contrastive Learning

26 Sep 2022arXiv:2209.12400archive 2025-07-28

Jiequan Cui, Zhisheng Zhong, Zhuotao Tian, Shu Liu, Bei Yu, Jiaya Jia

In this paper, we propose the Generalized Parametric Contrastive Learning (GPaCo/PaCo) which works well on both imbalanced and balanced data. Based on theoretical analysis, we observe that supervised contrastive loss tends to bias high-frequency classes and thus increases the difficulty of imbalanced learning. We introduce a set of parametric class-wise learnable centers to rebalance from an optimization perspective. Further, we analyze our GPaCo/PaCo loss under a balanced setting. Our analysis demonstrates that GPaCo/PaCo can adaptively enhance the intensity of pushing samples of the same class close as more samples are pulled together with their corresponding centers and benefit hard example learning. Experiments on long-tailed benchmarks manifest the new state-of-the-art for long-tailed recognition. On full ImageNet, models from CNNs to vision transformers trained with GPaCo loss show better generalization performance and stronger robustness compared with MAE models. Moreover, GPaCo can be applied to the semantic segmentation task and obvious improvements are observed on the 4 most popular benchmarks. Our code is available at https://github.com/dvlab-research/Parametric-Contrastive-Learning.

PaperPDFCode

In Syntology View this paper on Syntology, its page in Syntology's graph. That page lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

dvlab-research/parametric-contrastive-learning officialmentioned in papermentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Classes Are Not Equal: An Empirical Study on Image Recognition Fairness (arXiv:2402.18133; that paper's own run record: 3 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
dvlab-research/imbalanced-learning mentioned on GitHubpytorchSyntology: not harvested report
dvlab-research/rescom mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Parametric Contrastive Learning (arXiv:2107.12028; that paper's own run record: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified; the one sample that ran constructed an object rather than computing a result); nothing checks that they implement this paper's method. report
jiequancui/Parametric-Contrastive-Learning mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Parametric Contrastive Learning (arXiv:2107.12028; that paper's own run record: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified; the one sample that ran constructed an object rather than computing a result); nothing checks that they implement this paper's method. 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

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

Tasks

Contrastive LearningDomain GeneralizationImage ClassificationLong-tail LearningSemantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Domain Generalization ImageNet-C GPaCo (ViT-L) mean Corruption Error (mCE) 39.0 #13 of 47 Archive leaderboard report
Domain Generalization ImageNet-R GPaCo (ViT-L) Top-1 Error Rate 39.7 #15 of 39 Archive leaderboard report
Domain Generalization ImageNet-Sketch GPaCo (ViT-L) Top-1 accuracy 48.3 #12 of 20 Archive leaderboard report
Image Classification ImageNet GPaCo (ViT-L) Top 1 Accuracy 86.01% #176 of 1060 Archive leaderboard report
Image Classification ImageNet GPaCo (Vit-B) Top 1 Accuracy 84.0% #361 of 1060 Archive leaderboard report
Image Classification ImageNet GPaCo (ResNet-50) Top 1 Accuracy 79.7% #745 of 1060 Archive leaderboard report
Image Classification iNaturalist 2018 GPaCo (ResNet-152) Top-1 Accuracy 78.1% #21 of 60 Archive leaderboard report
Image Classification iNaturalist 2018 GPaCo (ResNet-50) Top-1 Accuracy 75.4% #25 of 60 Archive leaderboard report
Long-tail Learning ImageNet-LT GPaCo (2-ResNeXt101-32x4d) Top-1 Accuracy 63.2 #11 of 69 Archive leaderboard report
Long-tail Learning Places-LT GPaCo (ResNet-152) Top-1 Accuracy 41.7 #13 of 29 Archive leaderboard report
Long-tail Learning iNaturalist 2018 GPaCo (2-R152) Top-1 Accuracy 79.8% #7 of 43 Archive leaderboard report
Long-tail Learning iNaturalist 2018 GPaCo (ResNet-152) Top-1 Accuracy 78.1% #8 of 43 Archive leaderboard report
Long-tail Learning iNaturalist 2018 GPaCo (ResNet-50) Top-1 Accuracy 75.4% #12 of 43 Archive leaderboard report
Semantic Segmentation ADE20K GPaCo (Swin-L) Validation mIoU 54.3 #64 of 235 Archive leaderboard report
Semantic Segmentation PASCAL Context GPaCo (ResNet101) mIoU 56.2 #25 of 66 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

Contrastive LearningMAESupervised Contrastive Loss

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