Density Estimation On Bsds300
Metrics
CD
EMD
Log-likelihood
MMD-CD
MMD-EMD
NLL
Results
Performance results of various models on this benchmark
| Paper Title | |||||||
|---|---|---|---|---|---|---|---|
| TAN | - | - | 159.8 | - | - | - | Transformation Autoregressive Networks |
| FFJORD | 0.683 | 0.0281 | 157.4 | 0.548 | 0.0227 | -157.4 | FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models |
| B-NAF | - | - | 157.36 | - | - | - | Block Neural Autoregressive Flow |
| MADE MoG | - | - | 153.71 | - | - | - | Masked Autoregressive Flow for Density Estimation |
| PaddingFlow | 0.495 | 0.0248 | - | 0.48 | 0.0212 | - | PaddingFlow: Improving Normalizing Flows with Padding-Dimensional Noise |
0 of 5 row(s) selected.