JS divergence used in the original GAN. Essentially, the Wasserstein metric is said to be weak as it requires the use of a weaker topology, thus making it ...
確定! 回上一頁