Personalized Image Generation on DreamBooth
Metric: Concept Preservation (CP) (higher is better)
Metric: Concept Preservation (CP) (higher is better)
| # | Model↕ | Concept Preservation (CP)▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | IP-Adapter-Plus ViT-H SDXL v1.0 | 0.833 | No | IP-Adapter: Text Compatible Image Prompt Adapter... | 2023-08-13 | Code |
| 2 | DreamBooth LoRA SDXL v1.0 | 0.598 | No | DreamBooth: Fine Tuning Text-to-Image Diffusion ... | 2022-08-25 | Code |
| 3 | IP-Adapter ViT-G SDXL v1.0 | 0.593 | No | IP-Adapter: Text Compatible Image Prompt Adapter... | 2023-08-13 | Code |
| 4 | BLIP-Diffusion SD v1.5 | 0.547 | No | BLIP-Diffusion: Pre-trained Subject Representati... | 2023-05-24 | Code |
| 5 | Emu2 SDXL v1.0 | 0.528 | No | Generative Multimodal Models are In-Context Lear... | 2023-12-20 | Code |
| 6 | DreamBooth SD v1.5 | 0.494 | No | DreamBooth: Fine Tuning Text-to-Image Diffusion ... | 2022-08-25 | Code |
| 7 | Textual Inversion SD v1.5 | 0.378 | No | An Image is Worth One Word: Personalizing Text-t... | 2022-08-02 | Code |