TTA datasets

	A	B	C	D	E	F	G
1	Task	Datasets	# domains	# size	# classes	Description	Comments
2	Image Classification	CIFAR10-C	15	150,000	10	Gaussian Noise, Defocus Blur, Fog, Contrast, etc.
3		CIFAR100-C		150,000	100
4		ImageNet-C		750,000	1,000
5		TinyImageNet-C		150,000	200
6		ImageNet-A	-	7,500	200	Used for ImageNet Pretrained Models
7		ImageNet-R	-	30,000	200	Used for ImageNet Pretrained Models
8		DomainNet-126	4	145,145	126	Real, Clipart, Painting, Sketch	subset of DomainNet
9		DomainNet	6	600,000	345	Clipart, Real, Sketch, Infograph, Painting, Quickdraw
10		Office-Home	4	15,500	65	Artistic, Clip Art, Product, Real World
11		Office	3	4,110	31	Amazon, DSLR, Webcam
12		VisDA-C	2	207,000	12	Synthetic, Real
13		Digits-Five	5	215,695	10	MNIST, SVHN, USPS, MNIST-M, SyntheticDigits
14		VLCS	4	10,729	5	Caltech101, LabelMe, SUN09, VOC2007
15		PACS	4	9,991	7	Artpainting, Cartoon ,Photo ,Sketch
16		Synthia	-	9,400
17		Cityscapes	-	3,475
18		Citiscapes-to-ACDC	4	4006		Fog, Nighttime, Rain, now
19
20	Object Detection	BDD100k	-	100,000	-	Used for each other
21		Foggy Cityscapes		3,475
22		KITTI		7,481
23		Sim10k		10,000
24		Cityscapes		3,475
25	Graph Classification	ENZYMES		600	6	Train/Val/Test	TUDataset: A collection of benchmark datasets for learning with graphs.
26		PROTEINS		1113	2	Train/Val/Test
27	Node Classification	Cora		2703	10	Train/Val/Test	Revisiting Semi-Supervised Learning with Graph Embeddings
28		Amz-Photo		7,650	10	Train/Val/Test	Pitfalls of graph neural network evaluation
29	Text Classification	Amazon-Feature	4	16,000	2	Book, DVD, Electronics, Kitchen	https://jmcauley.ucsd.edu/data/amazon/
30		Amazon-Text	4	80,000	2	Book, DVD, Electronics, Kitchen	Discriminative Feature Adaptation via Conditional Mean Discrepancy for Cross-Domain Text Classification
31	Speech Recognition	LibriSpeech	-	-	-	Used for each other	Librispeech: An ASR corpus based on public domain audio books
32		CHiME-3					The third 'chime'speech separation and recognition challenge: Analysis and outcomes
33		Common voice					Common voice: A massively-multilingual speech corpus
34		TEDLIUM-v3					Ted-lium 3: twice as much data and corpus repartition for experiments on speaker adaptation
35	Face Recognition	WebFace		494,414	10,575	WebFace to AR face	Learning face representation from scratch
36		AR face		2,600	100	WebFace to AR face	The ar face database: Cvc technical report, 24
37	Face Anti-Spoofing	Face datasets	4	7,130	-	Replay-Attack, OULU-NPU, CASIA-MFSD, MSU-MFSD	On the effectiveness of local binary patterns in face anti-spoofing
38							OULU-NPU: A mobile face presentation attack database with real-world variations
39							A face antispoofing database with diverse attacks
40							Face spoof detection with image distortion analysis
41
42	Clip-based Adaptaion	Recognition Datasets	15	-	-	ImageNet, ImageNet-A, ImageNet-V2, ImageNet-R, ImageNet-Sketch, Flower102, DTD, Pets, Cars, UCF101, Caltech101, Food101, SUN397, Aircraft, EuroSAT	Refer to https://arxiv.org/pdf/2209.07511.pdf
43	Clip-based Adaptaion	ImageNet	-	1.33M	1,000	Used as target domains	Refer to https://arxiv.org/pdf/2109.01134.pdf
44		Caltech101		8,242	100
45		OxfordPets		7,349	37
46		StanfordCars		16,185	196
47		Flowers102		8,189	102
48		Food101		101,000	101
49		FGVCAircraft		10,000	100
50		SUN397		39,700	397
51		DTD		5,640	47
52		EuroSAT		27,000	10
53		UCF101		13,320	101
54		ImageNetV2		10,000	1,000
55		ImageNet-Sketch		50,889	1,000
56		ImageNet-A		7,500	200
57		ImageNet-R		30,000	200
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100