2
$\begingroup$

I am trying to train a Generative Adversarial Network and ran the training a few times with same dataset and same parameters but it seems tp produce different results. Why this may happen?

$\endgroup$

1 Answer 1

2
$\begingroup$

The loss function error surfaces of both discriminator and generator are complex. There are many different optima on both surfaces that can yield usable solutions. There is a element of randomness where on the loss function error surfaces a training run with stop.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.