Joint class generation on CelebA-BMS (test)
91.57AccuracyAC-GAN
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| AC-GANOracle=true, Backbone=BigGAN2022.06 | 91.57 | 10.41 | 2.4 | 0.79 | 0.49 | 1.18 | 0.86 | |
| cGAN-PD + S2M samplingBackbone=BigGAN, MCMC iterations=2002022.06 | 90.44 | 10.63 | 2.46 | 0.84 | 0.48 | 1.4 | 0.88 | |
| CP-GANBackbone=BigGAN2022.06 | 87.36 | 13.38 | 2.43 | 0.78 | 0.45 | 1.08 | 0.82 | |
| GAN + S2M samplingBackbone=BigGAN, MCMC iterations=2002022.06 | 85.22 | 10.5 | 2.51 | 0.83 | 0.48 | 1.36 | 0.89 | |
| cGAN-PDOracle=true, Backbone=BigGAN2022.06 | 78.35 | 10.38 | 2.45 | 0.79 | 0.48 | 1.12 | 0.86 | |
| AC-GAN*Oracle=false, Backbone=BigGAN2022.06 | 52.01 | 12.39 | 2.32 | 0.79 | 0.49 | 1.17 | 0.85 | |
| cGAN-PD*Oracle=false, Backbone=BigGAN2022.06 | 39.84 | 10.97 | 2.32 | 0.79 | 0.48 | 1.19 | 0.86 |