Visual Combination Generation on Visual Combinations Dataset
0.162Copy RateQwen-Image
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen-Imagefull_name=Qwen-Image-2511, prompt=Combine the two images into a novel and non-trivial image inspired by them., images_per_method=3962026.02 | 0.162 | — | 0.189 | 0.106 | |
| Nano Bananaprompt=Combine the two images into a novel and non-trivial image inspired by them., images_per_method=3962026.02 | 0.091 | — | 0.197 | 0.003 | |
| Flux.1 Kontextprompt=constant input prompt used in training, images_per_method=3962026.02 | 0.028 | — | 0.003 | 0.854 | |
| Inspiration Seedsimages_per_method=3962026.02 | 0.023 | — | 0 | 0.015 |