Semantic Placement on HSSD (val)
31.3HPCLIP-UNet
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CLIP-UNetTraining Data=LAION-SP pretraining + HSSD finetuning2024.01 | 31.3 | 18.5 | 24.9 | 35.3 | |
| LLM + DetectorPrior=LLM, Detector=Detic2024.01 | 29.8 | 22.1 | 41 | 38.2 | |
| CLIP-UNetTraining Data=HSSD only2024.01 | 23 | 16.2 | 26.6 | 36.5 | |
| GPT4V2024.01 | 8.3 | — | — | — | |
| LLaVA2024.01 | 6.8 | 0 | 26.3 | 43.4 |