HyperGAN
Search…
Least Squares Loss
1
a,b,c = (config.labels or [-1,1,1])
2
d_loss = 0.5*((d_real - b)**2) + 0.5*((d_fake - a)**2)
3
g_loss = 0.5*((d_s
Copied!

examples

1
{
2
"class": "function:hypergan.losses.least_squares_loss.LeastSquaresLoss"
3
}
Copied!

options

attribute
description
type
labels
[a,b,c]. Defaults to [-1,1,1]
array of floats (optional)
Copy link