Panoptic Segmentation on ADE20K open-vocabulary
23.5PQODISE
Evaluation results
| Method | Links | |
|---|---|---|
| ODISETraining Data=COCO, Backbone=ViT-H2023.12 | 23.5 | |
| OpenSDTraining Data=COCO, Backbone=Swin-L2023.12 | 23.1 | |
| OpenSDTraining Data=COCO+O365, Backbone=R502023.12 | 23 | |
| OpenSDTraining Data=COCO+O365, Backbone=Swin-T2023.12 | 22.2 | |
| X-DecoderTraining Data=COCO+ITP, Backbone=Swin-L2023.12 | 21.8 | |
| OpenSeeDTraining Data=COCO+O365, Backbone=Swin-T2023.12 | 19.8 | |
| OpenSDTraining Data=COCO, Backbone=Swin-T2023.12 | 19.6 | |
| OpenSDTraining Data=COCO, Backbone=R502023.12 | 19.4 | |
| X-DecoderTraining Data=COCO+ITP, Backbone=Swin-T2023.12 | 18.8 | |
| HIPIETraining Data=COCO+O365, Backbone=R502023.12 | 18.4 | |
| OPSNetTraining Data=COCO+IN, Backbone=R502023.12 | 17.7 | |
| OpenSeeDTraining Data=COCO, Backbone=Swin-L2023.12 | 17.1 | |
| OpenSeeDTraining Data=COCO, Backbone=Swin-T2023.12 | 16 | |
| MaskCLIPTraining Data=COCO, Backbone=R502023.12 | 15.1 |