GAN Generator Input Code Demo - Normally Distributed Random Noise
text
GAN Generator Input Code Demo - Normally Distributed Random Noise
As we now understand, the generator of a GAN accepts random noise vectors as input, which become upsampled via transposed convolutions to ultimately become images.
Specifically, we discussed how the values of the input noise vectors are randomly sampled from a normal distribution. We'll elaborate more on this point now to understand its importance and demonstrate what this looks like in code.
quiz
resources
updates
Committed by on