Text-to-image generation on CUB
9.78FIDSwinv2-Imagen
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Swinv2-Imagen2022.10 | 9.78 | 8.44 | — | |
| LAFITE2021.11 | 10.48 | 5.97 | — | |
| LAFITE2022.10 | 10.48 | 5.97 | — | |
| DF-GAN2021.11 | 14.81 | 5.1 | — | |
| DM-GAN2021.11 | 16.09 | 4.75 | — | |
| DM-GAN2022.10 | 16.09 | 4.75 | — | |
| PCCM-GAN2022.10 | 22.15 | 4.65 | — | |
| AttnGAN2021.11 | 23.98 | 4.36 | — | |
| Ours (EBM)Sampling=0.08, NFE=31, Parameter Overhead=2M2026.05 | 25.98 | — | 0.282 | |
| Ours (Gen)Sampling=0.001, NFE=12026.05 | 26.15 | — | 0.278 | |
| Diff-CMVAESampling=38.12, NFE=251, Parameter Overhead=14M2026.05 | 28 | — | 0.272 | |
| DALL-E2022.10 | 56.1 | — | — | |
| CMVAESampling=0.001, NFE=12026.05 | 155.11 | — | 0.26 | |
| MMVAESampling=0.001, NFE=12026.05 | 232.2 | — | 0.242 | |
| SDGAN2022.10 | — | 4.64 | — |