Image Classification on WebVision-1000

16개 결과 · ⬇ CSV · JSON

Top-1 Accuracy

Rank	Model	Top-1 Accuracy	Top-5 Accuracy	ImageNet Top-1 Accuracy	ImageNet Top-5 Accuracy	Extra Training Data	Paper	Code	Year
1	MAM (ViT-B/16)	83.6	–	–	–	✓	Improving Image Recognition by Retrieving from Web-Scale Image-Text Data		2023
2	CurriculumNet (InceptionResNet-v2)	79.3%	93.6%	–	–		CurriculumNet: Weakly Supervised Learning from Large-Scale Web Images	MalongTech/CurriculumNet · guoshengcv/CurriculumNet	2018
3	ERM + CMW-Net-SL (InceptionResNet-V2)	77.9%	92.6%	–	–		CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning	xjtushujun/cmw-net	2022
4	NCR+Mixup+DA (ResNet-50)	76.8	–	–	–		Learning with Neighbor Consistency for Noisy Labels	google-research/scenic	2022
5	Heteroscedastic (InceptionResNet-v2)	76.6%	92.1%	–	–		Correlated Input-Dependent Label Noise in Large-Scale Image Classification		2021
6	MILe (ResNet-50-D)	76.5%	90.9%	68.7	86.4		Multi-label Iterated Learning for Image Classification with Label Ambiguity		2021
7	SCC (ResNet50-D)	75.78%	91.07%	70.66%	88.46%		Webly Supervised Image Classification with Self-Contained Confidence	bigvideoresearch/SCC · bigvideoresearch/Enigma · yuleiqin/fopro · +1	2020
8	NCR (ResNet-50)	75.7%	–	–	–		Learning with Neighbor Consistency for Noisy Labels	google-research/scenic	2022
9	VSGraph-LC (ResNet50)	75.48%	90.15%	69.42%	87.29%		Webly Supervised Image Classification with Metadata: Automatic Noisy Label Correction via Visual-Semantic Graph	bigvideoresearch/Enigma	2020
10	MILe (ResNet-50)	75.2%	90.3%	67.1	85.6		Multi-label Iterated Learning for Image Classification with Label Ambiguity		2021
11	HAR (InceptionResNet-v2)	75.0%	90.6%	67.1%	86.7%		Heteroskedastic and Imbalanced Deep Learning with Adaptive Regularization	kaidic/HAR	2020
12	MentorMix (InceptionResNet-V2)	74.3%	90.5%	67.5%	87.2%		Beyond Synthetic Noise: Deep Learning on Controlled Noisy Labels	google-research/google-research · ljy-hy/mentormix_pytorch · MindCode-4/code-7	2019
13	MoPro (ResNet-50)	73.9%	90.0%	67.8%	87.0%		MoPro: Webly Supervised Learning with Momentum Prototypes	salesforce/MoPro · yuleiqin/capro	2020
14	SOMNet (ResNet-50)	72.2%	89.5%	65.0%	85.1%		Learning from Web Data with Self-Organizing Memory Module		2019
15	CurriculumNet (Inception-v2）	72.1%	89.2%	64.8%	84.9%		CurriculumNet: Weakly Supervised Learning from Large-Scale Web Images	MalongTech/CurriculumNet · guoshengcv/CurriculumNet	2018
16	MentorNet (InceptionResNet-V2)	70.8%	88.0%	62.5%	83.0%		MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels	google/mentornet	2017