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

COCO dataset and benchmarks – Page 46 · SOTA2 Research

Research Datasets

COCO

Benchmarks

Task NameDataset NameSOTA MethodSort

Task NameDataset NameSOTA ResultTrendResults

Semantic Instance Segmentation

COCO20k single-object setting

14.6AP^mk@50

4

Class-Agnostic Instance Segmentation

COCO 20k

6.8APmk50

4

Short-caption Image-to-Text Retrieval

COCO

76.1R@1

4

Short-caption Text-to-Image Retrieval

COCO

56.4Recall@1

4

Multimodal Understanding

COCO

35.9Accuracy

4

Object Detection

COCO LoCo (40-20)

50.9AP

4

Object Detection

COCO LoCo (70+10)

50.7AP

4

Object Detection

COCO LoCo (40+40)

52.2AP

4

Compositional generation

COCO

0.03KID

4

COCO

0KID (x10³)

4

Object Detection

COCO (test)

52.17Texts-1 Score

4

Text-to-image upsampling

COCO 5k (test)

24.07FID

4

COCO (val)

62.7TR@1

4

Bounding Box Classification

COCO (val)

56Top-1 Acc

4

Image generation

COCO

34.54IS

4

Human Body Pose Estimation

COCO

81PCK@0.05

4

Attribute Presence Measurement

COCO

0.933Attribute Presence

4

Instance Segmentation

COCO 2017 (val)

79.9Mean IoU

4

COCO 30K (val)

3.64NER (Threshold 0.3)

4

Panoptic Segmentation

COCO-C level 1 (val)

42.9PQ (Brightness)

4

Panoptic Segmentation

COCO-C level 2 corruptions (val)

42.2PQ (Brightness)

4

Image Captioning

COCO

27.6BLEU@4

4

COCO (val)

25.1CLIP Score

4

Interactive Object Detection

COCO (val)

68.4AP

4

Amodal Completion

COCO-A (test)

0Complete Failure Proportion

4

Page 46 of 57

Popular tasks

Object Detection Image Captioning Instance Segmentation Semantic segmentation Text-to-Image Generation

Follow for updates

Join SOTA2 Contact