List your company

ResearchFormerly WizWand

COCO

Benchmarks

TaskDatasetSOTA resultTrendResults

Image-to-Text Retrieval

COCO (test)

76.2R@1

26

Image Captioning

COCO (test)

149.2CIDEr

26

Grounded Text-to-Image Generation

COCO 2014 (val)

3.2FID

26

Semantic Segmentation

COCO-20i 5-shot

49.6Fold 0 Score

26

Object Detection

COCO-C

42mAP

26

Incremental Semantic Segmentation

COCO

50.4mIoU (Classes 1-60)

26

Multi-label Image Classification

MS-COCO

90.3mAP

26

Caption Retrieval

MS COCO Karpathy 5k (test)

50.1R@1

26

Image Captioning

COCO

15.8COCO Score

25

Whole-body pose estimation

COCO-WholeBody (val)

74.4Whole AP

25

Semantic Segmentation

COCO

83.41mIoU

25

Open-Vocabulary Segmentation

COCO-Object (COCO-O) (val)

67.9mIoU

25

Object Detection

COCO 70+10 scenario 2017 (val)

55.2AP

25

Object Detection

COCO 40+40 scenario 2017 (val)

54.7AP

25

Object Detection

COCO open-vocabulary (test)

39.3Novel AP

25

Object Composition

COCO

14.6Recall@1

25

Object Detection

COCO

48.3mAP

25

Object affordance recognition

COCO 2017 (val)

90.84mAP

25

Sentence Retrieval

MS-COCO 1K (test)

80.2Recall@1

25

Object Detection

MS-COCO 2% labeled protocol

27.7mAP

25

Text-to-Image Generation

COCO 2014 (test)

3.22FID

25

Visual Question Answering

VIS-IR COCO Infrared

64.72Accuracy

24

Text-to-Image Generation

COCO truck concept 'a photo of a truck' prompt (test)

19.5BRISQUE

24

Promptable Image Segmentation

COCO (val)

74mIoU

24

Gender Classification

COCO 95% spurious correlation

78.1Average Score

24