Zero Shot Segmentation On Ade20K Training
评估指标
mIoU
评测结果
各个模型在此基准测试上的表现结果
| Paper Title | Repository | ||
|---|---|---|---|
| COSMOS ViT-B/16 | 17.7 | COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training | |
| GEM (MetaCLIP) | 17.1 | Grounding Everything: Emerging Localization Properties in Vision-Language Transformers | |
| GEM (CLIP) | 15.7 | Grounding Everything: Emerging Localization Properties in Vision-Language Transformers | |
| CLIPSurgery | 12.9 | A Closer Look at the Explainability of Contrastive Language-Image Pre-training | |
| MaskCLIP | 10.2 | Extract Free Dense Labels from CLIP |
0 of 5 row(s) selected.