TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
Long-tail Learning	CIFAR-100-LT (ρ=100)	PCL	Error Rate	49.10	# 25
Long-tail Learning	CIFAR-10-LT (ρ=10)	PCL	Error Rate	9.14	# 14
Image Classification	ImageNet	ResNet-152	Top 1 Accuracy	81.3%	# 595
Image Classification	ImageNet	ResNet-200	Top 1 Accuracy	81.8%	# 553
Image Classification	ImageNet	ResNet-101	Top 1 Accuracy	80.9%	# 618
Long-tail Learning	ImageNet-LT	PaCo(ResNeXt-50)	Top-1 Accuracy	58.2	# 19
Long-tail Learning	ImageNet-LT	PaCo(ResNeXt101-32x4d)	Top-1 Accuracy	60.0	# 14
Image Classification	iNaturalist 2018	PaCo(ResNet-152)	Top-1 Accuracy	75.2%	# 23
Long-tail Learning	iNaturalist 2018	PaCo(ResNet-152)	Top-1 Accuracy	75.2%	# 12
Long-tail Learning	Places-LT	PaCo	Top-1 Accuracy	41.2	# 14

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/long-tail-learning-on-inaturalist-2018)](https://paperswithcode.com/sota/long-tail-learning-on-inaturalist-2018?p=parametric-contrastive-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/long-tail-learning-on-cifar-10-lt-r-10)](https://paperswithcode.com/sota/long-tail-learning-on-cifar-10-lt-r-10?p=parametric-contrastive-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/long-tail-learning-on-imagenet-lt)](https://paperswithcode.com/sota/long-tail-learning-on-imagenet-lt?p=parametric-contrastive-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/long-tail-learning-on-places-lt)](https://paperswithcode.com/sota/long-tail-learning-on-places-lt?p=parametric-contrastive-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/image-classification-on-inaturalist-2018)](https://paperswithcode.com/sota/image-classification-on-inaturalist-2018?p=parametric-contrastive-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/long-tail-learning-on-cifar-100-lt-r-100)](https://paperswithcode.com/sota/long-tail-learning-on-cifar-100-lt-r-100?p=parametric-contrastive-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/parametric-contrastive-learning/image-classification-on-imagenet)](https://paperswithcode.com/sota/image-classification-on-imagenet?p=parametric-contrastive-learning)`

Parametric Contrastive Learning

ICCV 2021 · Jiequan Cui, Zhisheng Zhong, Shu Liu, Bei Yu, Jiaya Jia ·

In this paper, we propose Parametric Contrastive Learning (PaCo) to tackle long-tailed recognition. Based on theoretical analysis, we observe supervised contrastive loss tends to bias on 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 PaCo loss under a balanced setting. Our analysis demonstrates that 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 CIFAR, ImageNet, Places, and iNaturalist 2018 manifest the new state-of-the-art for long-tailed recognition. On full ImageNet, models trained with PaCo loss surpass supervised contrastive learning across various ResNet backbones, e.g., our ResNet-200 achieves 81.8% top-1 accuracy. Our code is available at https://github.com/dvlab-research/Parametric-Contrastive-Learning.

PDF Abstract ICCV 2021 PDF ICCV 2021 Abstract

Code

Add Remove Mark official

jiequancui/Parametric-Contrastive-L… official

221

dvlab-research/parametric-contrasti… official

221

dvlab-research/imbalanced-learning

dvlab-research/rescom

silicx/dlsa

Tasks

Add Remove

Contrastive Learning

Image Classification

Long-tail Learning

Datasets

CIFAR-10

ImageNet

CIFAR-100

iNaturalist ImageNet-LT Places-LT

Results from the Paper

Edit

Ranked #12 on Long-tail Learning on iNaturalist 2018

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
Long-tail Learning	CIFAR-100-LT (ρ=100)	PCL	Error Rate	49.10	# 25	Compare
Long-tail Learning	CIFAR-10-LT (ρ=10)	PCL	Error Rate	9.14	# 14	Compare
Image Classification	ImageNet	ResNet-152	Top 1 Accuracy	81.3%	# 595	Compare
Image Classification	ImageNet	ResNet-200	Top 1 Accuracy	81.8%	# 553	Compare
Image Classification	ImageNet	ResNet-101	Top 1 Accuracy	80.9%	# 618	Compare
Long-tail Learning	ImageNet-LT	PaCo(ResNeXt-50)	Top-1 Accuracy	58.2	# 19	Compare
Long-tail Learning	ImageNet-LT	PaCo(ResNeXt101-32x4d)	Top-1 Accuracy	60.0	# 14	Compare
Image Classification	iNaturalist 2018	PaCo(ResNet-152)	Top-1 Accuracy	75.2%	# 23	Compare
Long-tail Learning	iNaturalist 2018	PaCo(ResNet-152)	Top-1 Accuracy	75.2%	# 12	Compare
Long-tail Learning	Places-LT	PaCo	Top-1 Accuracy	41.2	# 14	Compare

Methods

Add Remove

1x1 Convolution • Average Pooling • Batch Normalization • Bottleneck Residual Block • Contrastive Learning • Convolution • Global Average Pooling • Kaiming Initialization • Max Pooling • ReLU • Residual Block • Residual Connection • ResNet • Supervised Contrastive Loss

Edit Social Preview

Parametric Contrastive Learning

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit

Methods Edit Add Remove

Code

Add Remove Mark official

Tasks

Add Remove

Datasets

Results from the Paper

Edit

Methods

Add Remove