Density Estimation On Mnist
Metrics
COV-L2
MMD-L2
NLL (bits/dim)
Results
Performance results of various models on this benchmark
Model Name | COV-L2 | MMD-L2 | NLL (bits/dim) | Paper Title | Repository |
|---|---|---|---|---|---|
| FFJORD | 96.4% | 17.3 | 0.99 | FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models | |
| PaddingFlow | 100% | 11.0 | - | PaddingFlow: Improving Normalizing Flows with Padding-Dimensional Noise | |
| Identity | - | - | 0.134 | Backpropagation through Combinatorial Algorithms: Identity with Projection Works | |
| RNODE | - | - | 0.97 | How to train your neural ODE: the world of Jacobian and kinetic regularization | |
| MADE MoG | - | - | - | Masked Autoregressive Flow for Density Estimation |
0 of 5 row(s) selected.