Closed chenbofeng123 closed 4 years ago
kl_divergence = 0.5/ A_pred.size(0) (1 + 2model.logstd - model.mean2 - torch.exp(model.logstd)).sum(1).mean() torch.exp(model.logstd)2 ?
Dear @chenbofeng123, Thank you for your correction! I'll update it.
kl_divergence = 0.5/ A_pred.size(0) (1 + 2model.logstd - model.mean2 - torch.exp(model.logstd)).sum(1).mean() torch.exp(model.logstd)2 ?