Large Scale Adversarial Representation Learning

NeurIPS 2019 Jeff DonahueKaren Simonyan

Adversarially trained generative models (GANs) have recently achieved compelling image synthesis results. But despite early successes in using GANs for unsupervised representation learning, they have since been superseded by approaches based on self-supervision... (read more)

PDF Abstract

Evaluation Results from the Paper


 SOTA for Image Generation on ImageNet 64x64 (Inception Score metric )

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Self-Supervised Image Classification ImageNet BigBiGAN (RevNet-50 ×4, BN+CReLU) Top 1 Accuracy 61.3% # 15
Self-Supervised Image Classification ImageNet BigBiGAN (RevNet-50 ×4, BN+CReLU) Top 5 Accuracy 81.9% # 9
Self-Supervised Image Classification ImageNet BigBiGAN (RevNet-50 ×4, BN+CReLU) Number of Params 86M # 1
Self-Supervised Image Classification ImageNet BigBiGAN (ResNet-50) Top 1 Accuracy 55.4% # 19
Self-Supervised Image Classification ImageNet BigBiGAN (ResNet-50) Top 5 Accuracy 77.4% # 13
Self-Supervised Image Classification ImageNet BigBiGAN (ResNet-50, BN+CReLU) Top 1 Accuracy 56.6% # 18
Self-Supervised Image Classification ImageNet BigBiGAN (ResNet-50, BN+CReLU) Top 5 Accuracy 78.6% # 11
Self-Supervised Image Classification ImageNet BigBiGAN (RevNet-50 ×4) Top 1 Accuracy 60.8% # 16
Self-Supervised Image Classification ImageNet BigBiGAN (RevNet-50 ×4) Top 5 Accuracy 81.4% # 10
Semi-Supervised Image Classification ImageNet - 10% labeled data BigBiGAN (RevNet-50 ×4, BN+CReLU) Top 5 Accuracy 78.8% # 15
Semi-Supervised Image Classification ImageNet - 1% labeled data BigBiGAN (RevNet-50 ×4, BN+CReLU) Top 5 Accuracy 55.2% # 6
Image Generation ImageNet 64x64 BigBiGAN Inception Score 19.40 # 1
Image Generation ImageNet 64x64 BigBiGAN FID 15.82 # 1
Image Generation ImageNet 64x64 BigBiGAN Cls 47.51% # 1