GAN Training Dynamics

Hard

neural-networks DeepMind Meta NVIDIA

Multiple Choice

The generator produces only a few distinct outputs that fool the discriminator, ignoring the diversity of the real data distribution The discriminator becomes too weak and classifies all samples as real, halting learning Both networks converge to constant outputs, producing identical predictions for all inputs The generator's weights collapse to zero due to vanishing gradients from the discriminator

Correct!

Incorrect — try again next time!

Upgrade to Pro to unlock solutions and explanations