Attribute Binding in Text-to-Image Generation on BLIP-VQA
77.74Color BindingCoMat
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CoMatBase Model=SDXL, Train=true2024.11 | 77.74 | 65.91 | 52.62 | |
| ToMeBase Model=SDXL, Train=false2024.11 | 76.56 | 68.94 | 60.51 | |
| ELLABase Model=SDXL, Train=false2024.11 | 72.6 | 66.86 | 56.34 | |
| SynGenBase Model=SDXL, Train=false2024.11 | 70.1 | 60.44 | 50.69 | |
| ELLABase Model=SD1.5, Train=true2024.11 | 69.11 | 63.08 | 49.38 | |
| RanniBase Model=SDXL, Train=true2024.11 | 68.93 | 63.25 | 49.34 | |
| SynGenBase Model=SD1.5, Train=false2024.11 | 66.19 | 64.51 | 46.66 | |
| GORSBase Model=SD2.0, Train=false2024.11 | 66.03 | 62.87 | 47.85 | |
| CoMatBase Model=SD1.5, Train=true2024.11 | 65.61 | 61.9 | 49.75 | |
| Attn-Exct v2Base Model=SD2.0, Train=false2024.11 | 64 | 59.63 | 45.17 | |
| SDXLBase Model=-, Train=true2024.11 | 63.69 | 56.37 | 54.08 | |
| PlayG-v2Base Model=-, Train=true2024.11 | 62.08 | 61.25 | 50.87 | |
| DALL-E2Base Model=-, Train=true2024.11 | 57.5 | 63.74 | 54.64 | |
| SD v2Base Model=-, Train=true2024.11 | 50.65 | 49.22 | 42.21 | |
| Structured v2Base Model=SD2.0, Train=false2024.11 | 49.9 | 49 | 42.18 | |
| Composable v2Base Model=SD2.0, Train=false2024.11 | 40.63 | 36.45 | 32.99 | |
| SD v1.5Base Model=-, Train=true2024.11 | 37.5 | 41.59 | 37.24 | |
| RanniBase Model=SD1.5, Train=true2024.11 | 24.14 | 30.29 | 28.57 |