ZZUTK / Face-Aging-CAAE

Age Progression/Regression by Conditional Adversarial Autoencoder
636 stars 233 forks source link

Is it mode collapse? #45

Open chengxiqiao opened 4 years ago

chengxiqiao commented 4 years ago

I trained this code on CACD dataset and changed 10 labels to five. But the generated face look the same, it seems face aging doesn't work. Did you meet this problem? @ZZUTK

chengxiqiao commented 4 years ago

33

domitix commented 4 years ago

how did you manage to have these results? i mean, how many epochs etc?

chengxiqiao commented 4 years ago

how did you manage to have these results? i mean, how many epochs etc?

epoch is set 50, and the weights are (0.001,0,0) i didn't change @domitix