Text-to-Image Generation on HPSv2 (test)
8.019Aesthetic ScoreInPO-SD1.5
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| InPO-SD1.5baseline_comparison=Base-SD1.52025.03 | 8.019 | 0.9022 | — | 85.84 | — | 66.44 | — | — | — | — | |
| Semi-DPOBase=SD1.5, Compared Against=SD1.52026.04 | 7.88 | — | 0.798 | 0.874 | — | 55.8 | — | — | 0.881 | 73.7 | |
| Diff-KTOBase=SD1.5, Compared Against=SD1.52026.04 | 7.22 | — | 0.778 | 0.741 | — | 55.3 | — | — | 0.904 | 68.6 | |
| InPO-SD1.5baseline_comparison=DPO-SD1.52025.03 | 7.05 | 0.8406 | — | 74.94 | — | 61.09 | — | — | — | — | |
| Semi-DPOBase=SD1.5, Compared Against=Diff-DPO2026.04 | 6.84 | — | 0.746 | 0.774 | — | 52.2 | — | — | 0.83 | 66.4 | |
| Diff-DPOBase=SD1.5, Compared Against=SD1.52026.04 | 6.81 | — | 0.608 | 0.757 | — | 54.7 | — | — | 0.701 | 67.2 | |
| MaPOBase=SDXL, Compared Against=SDXL2026.04 | 6.71 | — | 0.646 | 0.558 | — | 55.4 | — | — | 0.789 | 56.1 | |
| InPO-SDXLbaseline_comparison=SFT-SDXL2025.03 | 6.666 | 0.8991 | — | 88.31 | — | 59 | — | — | — | — | |
| SPIN-DiffusionIteration=22024.02 | 6.3401 | 0.2839 | 1.2711 | 22.2577 | 7.5382 | — | — | — | — | — | |
| SPIN-DiffusionIteration=32024.02 | 6.3296 | 0.2849 | 1.2853 | 22.3029 | 7.5507 | — | — | — | — | — | |
| Semi-DPOBase=SDXL, Compared Against=SDXL2026.04 | 6.25 | — | 0.689 | 0.604 | — | 52.4 | — | — | 0.829 | 64.7 | |
| ArenaPOBackbone=Stable Diffusion XL2026.05 | 6.1696 | 0.2954 | 1.0463 | 0.2328 | — | 38.94 | — | — | — | — | |
| SDPOBackbone=Stable Diffusion XL2026.05 | 6.1572 | 0.2943 | 1.0363 | 0.2315 | — | 38.86 | — | — | — | — | |
| SPIN-DiffusionIteration=12024.02 | 6.1519 | 0.2803 | 1.1331 | 21.858 | 7.3558 | — | — | — | — | — | |
| Diff.-DPOBackbone=Stable Diffusion XL2026.05 | 6.1418 | 0.2941 | 1.0192 | 0.2298 | — | 38.71 | — | — | — | — | |
| DSPOBackbone=Stable Diffusion XL2026.05 | 6.1397 | 0.293 | 1.0148 | 0.2309 | — | 38.67 | — | — | — | — | |
| Pre-TrainedBackbone=Stable Diffusion XL2026.05 | 6.1271 | 0.29 | 0.9047 | 0.229 | — | 38.47 | — | — | — | — | |
| Semi-DPOBase=SDXL, Compared Against=Diff-DPO2026.04 | 6.12 | — | 0.529 | 0.419 | — | 42.5 | — | — | 0.528 | 54.4 | |
| MaPOBackbone=Stable Diffusion XL2026.05 | 6.1197 | 0.2891 | 0.8934 | 0.2286 | — | 38.4 | — | — | — | — | |
| SFTBackbone=Stable Diffusion XL2026.05 | 6.1084 | 0.2879 | 0.8749 | 0.2267 | — | 38.23 | — | — | — | — | |
| Semi-DPOBase=SD1.5, Compared Against=Diff-KTO2026.04 | 6.06 | — | 0.561 | 0.729 | — | 51.2 | — | — | 0.528 | 61.7 | |
| SFT2024.02 | 6.0057 | 0.2847 | 1.308 | 21.8211 | 7.3549 | — | — | — | — | — | |
| Diffusion-DPO2024.02 | 5.9878 | 0.2847 | 1.3085 | 22.2854 | 7.4666 | — | — | — | — | — | |
| InPO-SD1.5baseline_comparison=KTO-SD1.52025.03 | 5.981 | 0.6094 | — | 67.41 | — | 60.03 | — | — | — | — | |
| ArenaPOBackbone=Stable Diffusion 1.52026.05 | 5.8993 | 0.2882 | 0.6974 | 0.2215 | — | 36.59 | — | — | — | — | |
| DSPOBackbone=Stable Diffusion 1.52026.05 | 5.8728 | 0.2844 | 0.6591 | 0.2194 | — | 36.02 | — | — | — | — | |
| SDPOBackbone=Stable Diffusion 1.52026.05 | 5.8706 | 0.2851 | 0.6628 | 0.2199 | — | 36.29 | — | — | — | — | |
| SD-1.52024.02 | 5.8529 | 0.2781 | 0.9324 | 21.4825 | 7.1365 | — | — | — | — | — | |
| InPO-SD1.5Backbone=SD1.5, Alignment=Interactive Preference Optimization2025.03 | 5.8056 | 0.2852 | — | 21.916 | — | 36.876 | 5.7734 | 21.894 | — | — | |
| Diff.-DPOBackbone=Stable Diffusion 1.52026.05 | 5.7837 | 0.2817 | 0.5837 | 0.2188 | — | 36.13 | — | — | — | — | |
| SFTBackbone=Stable Diffusion 1.52026.05 | 5.7636 | 0.2829 | 0.6078 | 0.2189 | — | 35.73 | — | — | — | — | |
| MaPOBackbone=Stable Diffusion 1.52026.05 | 5.7521 | 0.2792 | 0.5111 | 0.218 | — | 35.57 | — | — | — | — | |
| SFT-SD1.5Backbone=SD1.5, Alignment=Supervised Fine-Tuning2025.03 | 5.7515 | 0.2803 | — | 21.648 | — | 36.292 | 5.7255 | 21.647 | — | — | |
| KTO-SD1.5Backbone=SD1.5, Alignment=Kahneman-Tversky Optimization2025.03 | 5.7248 | 0.2838 | — | 21.582 | — | 35.902 | 5.6922 | 21.566 | — | — | |
| InPO-SD1.5baseline_comparison=SFT-SD1.52025.03 | 5.663 | 0.5891 | — | 66.84 | — | 57.87 | — | — | — | — | |
| InPO-SDXLbaseline_comparison=Base-SDXL2025.03 | 5.637 | 0.8538 | — | 79.25 | — | 53.37 | — | — | — | — | |
| InPO-SDXLbaseline_comparison=DPO-SDXL2025.03 | 5.587 | 0.6581 | — | 59.28 | — | 46.91 | — | — | — | — | |
| DPO-SD1.5Backbone=SD1.5, Alignment=Direct Preference Optimization2025.03 | 5.5841 | 0.2743 | — | 21.297 | — | 35.591 | 5.5219 | 21.274 | — | — | |
| TGOBackbone=SD v1.52026.05 | 5.53 | — | 0.7595 | 21.45 | — | 30.38 | — | — | 0.2961 | — | |
| SFTBackbone=SD v1.52026.05 | 5.51 | — | 0.4744 | 21.09 | — | 28.79 | — | — | 0.2733 | — | |
| Diffusion-KTOBackbone=SD v1.52026.05 | 5.5 | — | 0.7365 | 21.24 | — | 30.02 | — | — | 0.2873 | — | |
| Pre-TrainedBackbone=Stable Diffusion 1.52026.05 | 5.4933 | 0.2697 | -0.0469 | 0.2088 | — | 34.8 | — | — | — | — | |
| CSFTBackbone=SD v1.52026.05 | 5.46 | — | 0.3703 | 20.89 | — | 28.74 | — | — | 0.2654 | — | |
| Diff-DPOBase=SDXL, Compared Against=SDXL2026.04 | 5.43 | — | 0.691 | 0.699 | — | 59.7 | — | — | 0.809 | 61.9 | |
| Diffusion-DPOBackbone=SD v1.52026.05 | 5.39 | — | 0.3672 | 21.14 | — | 29.31 | — | — | 0.2593 | — | |
| Base-SD1.5Backbone=SD1.52025.03 | 5.3848 | 0.2665 | — | 20.727 | — | 34.276 | 5.3491 | 20.719 | — | — | |
| SD v1.5Backbone=SD v1.52026.05 | 5.29 | — | 0.1384 | 20.71 | — | 28.9 | — | — | 0.2455 | — | |
| AlignPropBackbone=SD v1.52026.05 | 5.29 | — | 0.1066 | 20.64 | — | 28.82 | — | — | 0.2412 | — | |
| DSPOBackbone=SD v1.52026.05 | 5.29 | — | 0.1066 | 20.64 | — | 28.82 | — | — | 0.2412 | — | |
| Semi-DPOBase=SDXL, Compared Against=MaPO2026.04 | 4.94 | — | 0.601 | 0.572 | — | 46.6 | — | — | 0.655 | 60.9 | |
| EffiLoRABase Model=Stable Diffusion v1.5, Fine-tuned Dataset=pokemon-blip-captions2025.11 | — | — | — | — | — | — | — | — | 0.248 | — | |
| HydraLoRABase Model=Stable Diffusion v1.5, Fine-tuned Dataset=pokemon-blip-captions2025.11 | — | — | — | — | — | — | — | — | 0.2429 | — | |
| LoRABase Model=Stable Diffusion v1.5, Fine-tuned Dataset=pokemon-blip-captions2025.11 | — | — | — | — | — | — | — | — | 0.2408 | — |