Revision as of 23:06, 6 May 2020 edit 2601:600:9d7f:5a00:b0dc:a42f:b800:82b2 (talk) →Generative Adversarial Networks (GAN) ← Previous edit		Revision as of 23:07, 6 May 2020 edit undo 2601:600:9d7f:5a00:b0dc:a42f:b800:82b2 (talk) →Generative Adversarial Networks (GAN) Next edit →
Line 108: === Generative Adversarial Networks (GAN) === VQ has been used to quantize the feature in the discriminator of GANs. The feature quantization (FQ) technique ~~perform~~performs implicit feature matching<ref>Feature Quantization Improves GAN Training https://arxiv.org/abs/2004.02088</ref>. It improves the GAN training, and yield improved performance on a variety of popular GAN models: BigGAN for image generation, StyleGAN for face synthesis, and U-GAT-IT for unsupervised image-to-image translation. == See also ==

Vector quantization: Difference between revisions