Personalized Image Generation On Dreambench
평가 지표
Concept Preservation (CP)
Overall (CP * PF)
Prompt Following (PF)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | Concept Preservation (CP) | Overall (CP * PF) | Prompt Following (PF) | Paper Title | Repository |
|---|---|---|---|---|---|
| DreamBooth SD v1.5 | 0.494 | 0.356 | 0.721 | DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation | |
| DreamBooth LoRA SDXL v1.0 | 0.598 | 0.517 | 0.865 | DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation | |
| IP-Adapter-Plus ViT-H SDXL v1.0 | 0.833 | 0.344 | 0.413 | IP-Adapter: Text Compatible Image Prompt Adapter for Text-to-Image Diffusion Models | |
| Emu2 SDXL v1.0 | 0.528 | 0.364 | 0.690 | Generative Multimodal Models are In-Context Learners | |
| Textual Inversion SD v1.5 | 0.378 | 0.236 | 0.624 | An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion | |
| IP-Adapter ViT-G SDXL v1.0 | 0.593 | 0.380 | 0.640 | IP-Adapter: Text Compatible Image Prompt Adapter for Text-to-Image Diffusion Models | |
| BLIP-Diffusion SD v1.5 | 0.547 | 0.271 | 0.495 | BLIP-Diffusion: Pre-trained Subject Representation for Controllable Text-to-Image Generation and Editing |
0 of 7 row(s) selected.