Open-vocabulary semantic segmentation on Model Efficiency Analysis
70.3Params (M)CAT-Seg
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CAT-SegHardware=Nvidia-L40 GPU2024.11 | 70.3 | 693.7 | 535.2 | 3,459.7 | |
| LSMSegHardware=Nvidia-L40 GPU, Variation=w/o SAM-L2024.11 | 70.3 | 446 | 362.8 | 2,122 | |
| LSMSegHardware=Nvidia-L40 GPU, Variation=w/ SAM-L2024.11 | 73.4 | 546 | 426 | 3,140.6 | |
| OVSegHardware=Nvidia-L40 GPU2024.11 | 102.8 | — | 2,000 | 19,345.6 | |
| ZegFormerHardware=Nvidia-L40 GPU2024.11 | 103.3 | 1,148.3 | 2,700 | 19,425.6 |