Unsupervised open-vocabulary semantic segmentation on Pascal Context (val)
43.5mIoUFreeDA (ViT-L)
Evaluation Results
| Method | Links | |
|---|---|---|
| FreeDA (ViT-L)PAMR=false, Training Dataset=COCO Captions, Total Parameters (M)=732, Trainable Parameters (M)=0, Similarity=Visual2024.04 | 43.5 | |
| FreeDA (ViT-B)PAMR=false, Training Dataset=COCO Captions, Total Parameters (M)=236.1, Trainable Parameters (M)=0, Similarity=Visual2024.04 | 43.1 | |
| FreeDA (ViT-L)PAMR=false2024.04 | 38.3 | |
| TCLPAMR=true, Training Dataset=CC3M+CC12M, Total Parameters (M)=178.3, Trainable Parameters (M)=21.7, Similarity=Visual2024.04 | 33.9 | |
| OVDiffPAMR=false, Total Parameters (M)=1226.4, Trainable Parameters (M)=0, Similarity=Visual2024.04 | 33.7 | |
| TCLPAMR=true, Training Dataset=CC3M+CC12M2024.04 | 30.4 | |
| TCLPAMR=false, Training Dataset=CC3M+CC12M, Total Parameters (M)=178.3, Trainable Parameters (M)=21.7, Similarity=Textual2024.04 | 30.3 | |
| MaskCLIPPAMR=false, Total Parameters (M)=291, Trainable Parameters (M)=0, Similarity=Textual2024.04 | 26.4 | |
| MaskCLIPPAMR=true, Total Parameters (M)=291, Trainable Parameters (M)=0, Similarity=Textual2024.04 | 25.3 | |
| SegCLIPPAMR=false, Training Dataset=CC3M+COCO Captions2024.04 | 24.7 | |
| ReCoPAMR=true, Training Dataset=ImageNet1k, Total Parameters (M)=313, Trainable Parameters (M)=0, Similarity=Visual2024.04 | 24.7 | |
| TCLPAMR=false, Training Dataset=CC3M+CC12M2024.04 | 24.3 | |
| GroupViTPAMR=true, Training Dataset=CC12M+YFCC, Total Parameters (M)=55.8, Trainable Parameters (M)=55.8, Similarity=Visual2024.04 | 23.8 | |
| MaskCLIPPAMR=false2024.04 | 23.6 | |
| GroupViTPAMR=false, Training Dataset=CC12M+RedCaps, Total Parameters (M)=55.8, Trainable Parameters (M)=55.8, Similarity=Textual2024.04 | 23.4 | |
| ViewCoPAMR=false, Training Dataset=CC12M+YFCC2024.04 | 23 | |
| MaskCLIPPAMR=true2024.04 | 22.6 | |
| ReCoPAMR=false, Training Dataset=ImageNet1k, Total Parameters (M)=313, Trainable Parameters (M)=0, Similarity=Visual2024.04 | 22.3 | |
| OVSegmentorPAMR=false, Training Dataset=CC4M2024.04 | 20.4 | |
| ReCoPAMR=false2024.04 | 19.9 | |
| GroupViTPAMR=true, Training Dataset=CC12M+YFCC2024.04 | 19 | |
| GroupViTPAMR=false, Training Dataset=CC12M+RedCaps2024.04 | 18.7 |