Domain Adaptation on ImageNet-C

Metric: Top 1 Accuracy (higher is better)

LeaderboardDataset

Loading chart...

Results

Hide augmentations

Sort:

#	Model↕	Top 1 Accuracy▼	Augmentations	Paper	Date↕	Code
1	FAN-L-Hybrid (IN-22k)	73.6	Yes	Understanding The Robustness in Vision Transform...	2022-04-26	Code
2	FAN-B-Hybrid (IN-22k)	70.5	Yes	Understanding The Robustness in Vision Transform...	2022-04-26	Code
3	ResNet-50 (PushPull-Conv) + PRIME	69.4	No	PushPull-Net: Inhibition-driven ResNet robust to...	2024-08-07	Code
4	FAN-L-Hybrid+STL	69.2	No	Fully Attentional Networks with Self-emerging To...	2024-01-08	Code
5	FAN-L-Hybrid	67.7	No	Understanding The Robustness in Vision Transform...	2022-04-26	Code
6	DiffAUD (ConvNeXt-Tiny)	64.3	Yes	-	-	Code
7	DiffAUD (Swin-Tiny)	61	Yes	-	-	Code
8	PRIME + DeepAugment (ResNet-50)	59.9	No	PRIME: A few primitives can boost robustness to ...	2021-12-27	Code
9	ViT-B/16-SAM	56.5	No	When Vision Transformers Outperform ResNets with...	2021-06-03	Code
10	PRIME with JSD (ResNet-50)	56.4	No	PRIME: A few primitives can boost robustness to ...	2021-12-27	Code
11	PRIME (ResNet-50)	55	No	PRIME: A few primitives can boost robustness to ...	2021-12-27	Code
12	ResNet-152x2-SAM	55	No	When Vision Transformers Outperform ResNets with...	2021-06-03	Code
13	DiffAUD (ResNet-50)	52.1	Yes	-	-	Code
14	Mixer-B/8-SAM	48.9	No	When Vision Transformers Outperform ResNets with...	2021-06-03	Code

#1FAN-L-Hybrid (IN-22k)SOTA
73.6
Top 1 Accuracy· Augmentations· 2022-04-26
Understanding The Robustness in Vision Transformers Code
#2FAN-B-Hybrid (IN-22k)
70.5
Top 1 Accuracy· Augmentations· 2022-04-26
Understanding The Robustness in Vision Transformers Code
#3ResNet-50 (PushPull-Conv) + PRIME
69.4
Top 1 Accuracy· 2024-08-07
PushPull-Net: Inhibition-driven ResNet robust to image corruptions Code
#4FAN-L-Hybrid+STL
69.2
Top 1 Accuracy· 2024-01-08
Fully Attentional Networks with Self-emerging Token Labeling Code
#5FAN-L-Hybrid
67.7
Top 1 Accuracy· 2022-04-26
Understanding The Robustness in Vision Transformers Code
#6DiffAUD (ConvNeXt-Tiny)
64.3
Top 1 Accuracy· Augmentations
No paperCode
#7DiffAUD (Swin-Tiny)
61
Top 1 Accuracy· Augmentations
No paperCode
#8PRIME + DeepAugment (ResNet-50)SOTA
59.9
Top 1 Accuracy· 2021-12-27
PRIME: A few primitives can boost robustness to common corruptions Code
#9ViT-B/16-SAMSOTA
56.5
Top 1 Accuracy· 2021-06-03
When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations Code
#10PRIME with JSD (ResNet-50)
56.4
Top 1 Accuracy· 2021-12-27
PRIME: A few primitives can boost robustness to common corruptions Code
#11PRIME (ResNet-50)
55
Top 1 Accuracy· 2021-12-27
PRIME: A few primitives can boost robustness to common corruptions Code
#12ResNet-152x2-SAM
55
Top 1 Accuracy· 2021-06-03
When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations Code
#13DiffAUD (ResNet-50)
52.1
Top 1 Accuracy· Augmentations
No paperCode
#14Mixer-B/8-SAM
48.9
Top 1 Accuracy· 2021-06-03
When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations Code