Image Classification on mini WebVision 1.0

Metric: ImageNet Top-5 Accuracy (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	ImageNet Top-5 Accuracy▼	Extra Data	Paper	Date↕	Code
1	RTE (Inception-ResNet-v2)	97.24	No	Robust Temporal Ensembling for Learning with Noi...	2021-09-29	-
2	PSSCL (130 epochs)	95.16	No	-	-	Code
3	PSSCL (120 epochs)	94.84	No	-	-	Code
4	CC	93.86	No	Centrality and Consistency: Two-Stage Clean Samp...	2022-07-29	Code
5	Robust LR	93.76	No	Two Wrongs Don't Make a Right: Combating Confirm...	2021-12-06	-
6	CMW-Net-SL+C2D	93.48	No	CMW-Net: Learning a Class-Aware Sample Weighting...	2022-02-11	Code
7	PGDF (Inception-ResNet-v2)	93.11	No	Sample Prior Guided Robust Model Learning to Sup...	2021-12-02	Code
8	Dynamic Loss (Inception-ResNet-v2)	93.08	No	Dynamic Loss For Robust Learning	2022-11-22	Code
9	Sel-CL+ (ResNet-18)	93.04	No	Selective-Supervised Contrastive Learning with N...	2022-03-08	Code
10	FaMUS	92.76	No	Faster Meta Update Strategy for Noise-Robust Dee...	2021-04-30	Code
11	CMW-Net-SL	92.52	No	CMW-Net: Learning a Class-Aware Sample Weighting...	2022-02-11	Code
12	CoDiM-Self (Inception-ResNet-v2)	92.48	No	CoDiM: Learning with Noisy Labels via Contrastiv...	2021-11-23	-
13	ROLT+ (Inception-ResNet-v2)	92.48	No	Robust Long-Tailed Learning under Label Noise	2021-08-26	-
14	TCL	92.4	No	Twin Contrastive Learning with Noisy Labels	2023-03-13	Code
15	BtR	92.2	No	Bootstrapping the Relationship Between Images an...	2022-10-17	Code
16	CAR	92.09	No	Confidence Adaptive Regularization for Deep Lear...	2021-08-18	-
17	CoDiM-Sup (Inception-ResNet-v2)	91.96	No	CoDiM: Learning with Noisy Labels via Contrastiv...	2021-11-23	-
18	SSR	91.76	No	SSR: An Efficient and Robust Framework for Learn...	2021-11-22	Code
19	DivideMix (Inception-ResNet-v2)	91.64	No	DivideMix: Learning with Noisy Labels as Semi-su...	2020-02-18	Code
20	NCT (Inception-ResNet-v2)	91.61	No	Noisy Concurrent Training for Efficient Learning...	2020-09-17	Code
21	GJS (ResNet-50)	91.27	No	Generalized Jensen-Shannon Divergence Loss for L...	2021-05-10	Code
22	MentorMix (Inception-ResNet-v2)	91.1	No	Beyond Synthetic Noise: Deep Learning on Control...	2019-11-21	Code
23	NGC (Inception-ResNet-v2)	91.04	No	NGC: A Unified Framework for Learning with Open-...	2021-08-25	-
24	ELR+ (Inception-ResNet-v2)	89.76	No	Early-Learning Regularization Prevents Memorizat...	2020-06-30	Code
25	Crust (Inception-ResNet-v2)	87.84	No	Coresets for Robust Training of Neural Networks ...	2020-11-15	-
26	ODD (Inception-ResNet-v2)	86.3	No	Robust and On-the-fly Dataset Denoising for Imag...	2020-03-24	-
27	MentorNet (Inception-ResNet-v2)	85.8	No	MentorNet: Learning Data-Driven Curriculum for V...	2017-12-14	Code
28	Iterative-CV (Inception-ResNet-v2)	85	No	Understanding and Utilizing Deep Neural Networks...	2019-05-13	Code
29	Co-teaching (Inception-ResNet-v2)	84.7	No	Co-teaching: Robust Training of Deep Neural Netw...	2018-04-18	Code
30	F-Correction (Inception-ResNet-v2)	82.36	No	Making Deep Neural Networks Robust to Label Nois...	2016-09-13	Code
31	D2L (Inception-ResNet-v2)	81.36	No	Dimensionality-Driven Learning with Noisy Labels	2018-06-07	Code

#1RTE (Inception-ResNet-v2)SOTA
97.24
ImageNet Top-5 Accuracy· 2021-09-29
Robust Temporal Ensembling for Learning with Noisy Labels
#2PSSCL (130 epochs)
95.16
ImageNet Top-5 Accuracy
No paperCode
#3PSSCL (120 epochs)
94.84
ImageNet Top-5 Accuracy
No paperCode
#4CC
93.86
ImageNet Top-5 Accuracy· 2022-07-29
Centrality and Consistency: Two-Stage Clean Samples Identification for Learning with Instance-Dependent Noisy Labels Code
#5Robust LR
93.76
ImageNet Top-5 Accuracy· 2021-12-06
Two Wrongs Don't Make a Right: Combating Confirmation Bias in Learning with Label Noise
#6CMW-Net-SL+C2D
93.48
ImageNet Top-5 Accuracy· 2022-02-11
CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning Code
#7PGDF (Inception-ResNet-v2)
93.11
ImageNet Top-5 Accuracy· 2021-12-02
Sample Prior Guided Robust Model Learning to Suppress Noisy Labels Code
#8Dynamic Loss (Inception-ResNet-v2)
93.08
ImageNet Top-5 Accuracy· 2022-11-22
Dynamic Loss For Robust Learning Code
#9Sel-CL+ (ResNet-18)
93.04
ImageNet Top-5 Accuracy· 2022-03-08
Selective-Supervised Contrastive Learning with Noisy Labels Code
#10FaMUSSOTA
92.76
ImageNet Top-5 Accuracy· 2021-04-30
Faster Meta Update Strategy for Noise-Robust Deep Learning Code
#11CMW-Net-SL
92.52
ImageNet Top-5 Accuracy· 2022-02-11
CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning Code
#12CoDiM-Self (Inception-ResNet-v2)
92.48
ImageNet Top-5 Accuracy· 2021-11-23
CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning
#13ROLT+ (Inception-ResNet-v2)
92.48
ImageNet Top-5 Accuracy· 2021-08-26
Robust Long-Tailed Learning under Label Noise
#14TCL
92.4
ImageNet Top-5 Accuracy· 2023-03-13
Twin Contrastive Learning with Noisy Labels Code
#15BtR
92.2
ImageNet Top-5 Accuracy· 2022-10-17
Bootstrapping the Relationship Between Images and Their Clean and Noisy Labels Code
#16CAR
92.09
ImageNet Top-5 Accuracy· 2021-08-18
Confidence Adaptive Regularization for Deep Learning with Noisy Labels
#17CoDiM-Sup (Inception-ResNet-v2)
91.96
ImageNet Top-5 Accuracy· 2021-11-23
CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning
#18SSR
91.76
ImageNet Top-5 Accuracy· 2021-11-22
SSR: An Efficient and Robust Framework for Learning with Unknown Label Noise Code
#19DivideMix (Inception-ResNet-v2)SOTA
91.64
ImageNet Top-5 Accuracy· 2020-02-18
DivideMix: Learning with Noisy Labels as Semi-supervised Learning Code
#20NCT (Inception-ResNet-v2)
91.61
ImageNet Top-5 Accuracy· 2020-09-17
Noisy Concurrent Training for Efficient Learning under Label Noise Code
#21GJS (ResNet-50)
91.27
ImageNet Top-5 Accuracy· 2021-05-10
Generalized Jensen-Shannon Divergence Loss for Learning with Noisy Labels Code
#22MentorMix (Inception-ResNet-v2)SOTA
91.1
ImageNet Top-5 Accuracy· 2019-11-21
Beyond Synthetic Noise: Deep Learning on Controlled Noisy Labels Code
#23NGC (Inception-ResNet-v2)
91.04
ImageNet Top-5 Accuracy· 2021-08-25
NGC: A Unified Framework for Learning with Open-World Noisy Data
#24ELR+ (Inception-ResNet-v2)
89.76
ImageNet Top-5 Accuracy· 2020-06-30
Early-Learning Regularization Prevents Memorization of Noisy Labels Code
#25Crust (Inception-ResNet-v2)
87.84
ImageNet Top-5 Accuracy· 2020-11-15
Coresets for Robust Training of Neural Networks against Noisy Labels
#26ODD (Inception-ResNet-v2)
86.3
ImageNet Top-5 Accuracy· 2020-03-24
Robust and On-the-fly Dataset Denoising for Image Classification
#27MentorNet (Inception-ResNet-v2)SOTA
85.8
ImageNet Top-5 Accuracy· 2017-12-14
MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels Code
#28Iterative-CV (Inception-ResNet-v2)
85
ImageNet Top-5 Accuracy· 2019-05-13
Understanding and Utilizing Deep Neural Networks Trained with Noisy Labels Code
#29Co-teaching (Inception-ResNet-v2)
84.7
ImageNet Top-5 Accuracy· 2018-04-18
Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labels Code
#30F-Correction (Inception-ResNet-v2)SOTA
82.36
ImageNet Top-5 Accuracy· 2016-09-13
Making Deep Neural Networks Robust to Label Noise: a Loss Correction Approach Code
#31D2L (Inception-ResNet-v2)
81.36
ImageNet Top-5 Accuracy· 2018-06-07
Dimensionality-Driven Learning with Noisy Labels Code