generator feature
dd0151081b1e80e93f1b8aaf0b684c18-Paper-Conference.pdf
Wespanthe distilled representation of the synthetic domain tothe real domain by also using real training data to remedy the mode collapse of GANs and boost the student network performance in a real domain.
Distilling Representations from GAN Generator via Squeeze and Span: Appendix A Implementation Details Pre-trained GAN Table 1 presents the details of pre-trained GANs employed in our experiments
Furthermore, we provide an in-depth analysis of the cross-layer generator feature similarity.
Distilling Representations from GAN Generator via Squeeze and Span Y u Yang
In particular, we present a novel "squeeze-and-span" technique to distill knowledge from a generator
Industry:
- Education (0.68)
- Information Technology > Security & Privacy (0.46)
Technology: