High-Fidelity Image Generation With Fewer Labels

6 Mar 2019Mario LucicMichael TschannenMarvin RitterXiaohua ZhaiOlivier BachemSylvain Gelly

Deep generative models are becoming a cornerstone of modern machine learning. Recent work on conditional generative adversarial networks has shown that learning complex, high-dimensional distributions over natural images is within reach... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Conditional Image Generation ImageNet 128x128 S3 GAN FID 7.7 # 2
Conditional Image Generation ImageNet 128x128 S3 GAN Inception score 83.1 # 3