Subject-driven image generation on Multi-subject bench
0.302CLIP-TSSR-Encoder
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SSR-EncoderType=Finetune-free2023.12 | 0.302 | 0.182 | 0.556 | |
| DreamboothType=Finetune-based2023.12 | 0.298 | 0.223 | 0.681 | |
| BLIP-DiffusionType=Finetune-free2023.12 | 0.287 | 0.198 | 0.514 | |
| Break-A-SceneType=Finetune-based2023.12 | 0.285 | 0.187 | 0.63 | |
| IP-adapterType=Finetune-free2023.12 | 0.272 | 0.201 | 0.442 | |
| ELITEType=Finetune-free2023.12 | 0.253 | 0.194 | 0.483 | |
| Reference onlyType=Finetune-free2023.12 | 0.242 | 0.195 | 0.434 | |
| Textual InversionType=Finetune-based2023.12 | 0.24 | 0.212 | 0.41 |