Logistic Loss

Adapted from http://stylegan.xyz/paper​
Source: /losses/logistic_loss.py​
1
d_loss = self.softplus(-d_real) + self.softplus(d_fake)
2
g_loss = self.softplus(-d_fake)
Copied!
examples
Configurations: /losses/logistic_loss/​
1
{                                                                                       
2
  "class": "function:hypergan.losses.logistic_loss.LogisticLoss"
3
}
Copied!
options
attribute
description
type
beta
​https://pytorch.org/docs/stable/_modules/torch/nn/modules/activation.html#Softplus​
float (optional)
threshold
​https://pytorch.org/docs/stable/_modules/torch/nn/modules/activation.html#Softplus​
float (optional)

Least Squares Loss

QP Loss

Last modified 1yr ago

Copy link

Contents

examples

options