Autoencoding on CIFAR-10 (val)
1.07MSEMSE+Spectral
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MSE+SpectralAutoencoder=BM-SHJ2025.04 | 1.07 | 0.612 | 0.329 | 197.3 | 0.105 | 25.93 | 5 | |
| MSE+L1Autoencoder=BM-SHJ2025.04 | 2.22 | 0.984 | 0.178 | 13.42 | 0.066 | 0.74 | 7.2 | |
| MSE+LaplacianAutoencoder=BM-SHJ2025.04 | 3.24 | 0.977 | 0.196 | 19.8 | 0.091 | 0.6 | 4.2 | |
| MSE+SSIMAutoencoder=BM-SHJ2025.04 | 3.42 | 0.977 | 0.216 | 17.03 | 0.099 | 0.93 | 6.7 | |
| MSEAutoencoder=BM-SHJ2025.04 | 3.46 | 0.976 | 0.213 | 17.64 | 0.1 | — | 7.4 | |
| MSE+TVAutoencoder=BM-SHJ2025.04 | 3.8 | 0.56 | 0.375 | 149 | 0.638 | 0.08 | 2.1 | |
| MSE+GradMagAutoencoder=BM-SHJ2025.04 | 7.59 | 0.94 | 0.214 | 48.53 | 0.117 | 0.6 | 6.2 | |
| MSE+FDSEAutoencoder=BM-SHJ2025.04 | 8.73 | 0.9937 | 0.141 | 9.895 | 0.028 | 0.54 | 7.7 |