1) Train the discriminator. - Sample a batch of random points in the latent space. - Turn the points into fake images via the "generator" model.
確定! 回上一頁