List your company
ResearchFormerly WizWand

Generative Modeling on MNIST real-valued (test)

3,402Log-likelihoodAdversarial Autoencoder
265.361,079.681,8942,708.32Nov 18, 2015
Updated 8mo ago

Evaluation results

MethodLinks
Adversarial AutoencoderParzen window samples=10K2015.113,402
GMMN + AEParzen window samples=10K2015.112,822
GANParzen window samples=10K2015.112,252
Deep GSNParzen window samples=10K2015.112,141.1
DBNParzen window samples=10K2015.111,382
Stacked CAEParzen window samples=10K2015.111,211.6
Adversarial AutoencoderParzen window samples=10M2015.11427
GANParzen window samples=10M2015.11386