Exploring the Limits of Weakly Supervised Pretraining

ECCV 2018 Dhruv MahajanRoss GirshickVignesh RamanathanKaiming HeManohar PaluriYixuan LiAshwin BharambeLaurens van der Maaten

State-of-the-art visual perception models for a wide range of tasks rely on supervised pretraining. ImageNet classification is the de facto pretraining task for these models... (read more)

PDF Abstract

Evaluation Results from the Paper


#6 best model for Image Classification on ImageNet (using extra training data)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
COMPARE
Image Classification ImageNet ResNeXt-101 32x8d Top 1 Accuracy 82.2% # 27
Image Classification ImageNet ResNeXt-101 32x8d Top 5 Accuracy 96.4% # 16
Image Classification ImageNet ResNeXt-101 32x8d Number of params 88M # 1
Image Classification ImageNet ResNeXt-101 32x48d Top 1 Accuracy 85.4% # 6
Image Classification ImageNet ResNeXt-101 32x48d Top 5 Accuracy 97.6% # 4
Image Classification ImageNet ResNeXt-101 32x48d Number of params 829M # 1
Image Classification ImageNet ResNeXt-101 32x32d Top 1 Accuracy 85.1% # 8
Image Classification ImageNet ResNeXt-101 32x32d Top 5 Accuracy 97.5% # 5
Image Classification ImageNet ResNeXt-101 32x32d Number of params 466M # 1
Image Classification ImageNet ResNeXt-101 32×16d Top 1 Accuracy 84.2% # 14
Image Classification ImageNet ResNeXt-101 32×16d Top 5 Accuracy 97.2% # 8
Image Classification ImageNet ResNeXt-101 32×16d Number of params 194M # 1