Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

CC3M dataset and benchmarks · SOTA2 Research

Research Datasets

CC3M

Benchmarks

Task NameDataset NameSOTA MethodSort

Task NameDataset NameSOTA ResultTrendResults

Image Generation

CC3M SDXL v1.0 (test)

33.49FID

27

Aesthetic Evaluation

CC3M SDXL 1.0 (test)

0.293HPS

27

Object Recognition

CC3M (test)

0.738Recall

21

Aesthetic Evaluation

CC3M

0.276HPS

15

Image Generation

CC3M (test)

36.96FID

15

Multimodal Understanding

CC3M IOD

100Accuracy

14

Malicious Prompt Detection

CC3M (IOD)

0FPR

14

Text-to-Image Retrieval

CC3M

45.7Recall

9

Image-to-Text Retrieval

CC3M

47.2Recall

9

Image Classification

CC3M

46.7Accuracy

9

Multi-Tag Selection

CC3M (test)

92.5Precision

9

Text-to-image generation

CC3M

6.06FID

7

Image-Text Retrieval

CC3M-595K-LLaVA 1

7.6IR Score

6

Conditional Inversion

CC3M first 500 images

25.24PSNR

6

Unconditional Inversion

CC3M first 500 images

24.41PSNR

6

Multi-Tag Selection

CC3M

0.883Precision

6

Backdoor Defense (Multimodal)

CC3M 60K subset 1.0 (test)

64.03Clean Accuracy (CA)

5

Vision-Language Compositional Evaluation

CC3M 50,000 random subset TripletData

92.25Text Score

4

Text-level Semantic Segmentation

CC3M (subset)

65.5Caption IoU

4

Object Recognition

CC3M

86.8Recall

3

Popular tasks

Image Generation object recognition Multi-Tag Selection Aesthetic Evaluation Image Classification

Follow for updates

Join SOTA2 Contact