Text-to-Image Generation on Internal Corpus FLUX.1
0.3468CLIP I-TAttend&Excite
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Attend&ExciteFine-tuning=true, Optimal λ=true2025.10 | 0.3468 | 0.232 | 0.5876 | 0.6382 | 23.3333 | 1.3806 | 2.3477 | |
| FOCUS(Ours)Fine-tuning=true, Optimal λ=true2025.10 | 0.3468 | 0.2328 | 0.578 | 0.6386 | 23.3278 | 1.3899 | 2.5881 | |
| CONFORMFine-tuning=true, Optimal λ=true2025.10 | 0.3458 | 0.2305 | 0.58 | 0.6369 | 23.3724 | 1.3631 | 1.9591 | |
| BaseFine-tuning=true, Optimal λ=true2025.10 | 0.3449 | 0.2271 | 0.5739 | 0.63 | 23.4234 | 1.297 | 0 | |
| Divide&BindFine-tuning=true, Optimal λ=true2025.10 | 0.3445 | 0.2296 | 0.5705 | 0.6246 | 23.1909 | 1.2269 | 0.2002 |