Glow: Generative Flow with Invertible 1x1 Convolutions

Benchmark Model Rank Results
density-estimation-on-imagenet-32x32-1Glow#4NLL (bits/dim): 4.09
image-generation-on-celeba-256x256Glow (Kingma and Dhariwal, 2018)#9bpd: 1.03
image-generation-on-celeba-hq-256x256GLOW#13FID: 68.93
image-generation-on-imagenet-32x32Glow (Kingma and Dhariwal, 2018)#28bpd: 4.09
image-generation-on-imagenet-64x64Glow (Kingma and Dhariwal, 2018)#49Bits per dim: 3.81
unsupervised-anomaly-detection-on-smapGlow#5F1: 86.05Precision: 87.40Recall: 84.93AUC: 91.55