Data Augmentation On Imagenet

Metrics

Accuracy (%)

Results

Performance results of various models on this benchmark

		Paper Title
DeiT-B (+MixPro)	82.9	MixPro: Data Augmentation with MaskMix and Progressive Attention Labeling for Vision Transformer
ResNet-200 (DeepAA)	81.32	Deep AutoAugment
DeiT-S (+MixPro)	81.3	MixPro: Data Augmentation with MaskMix and Progressive Attention Labeling for Vision Transformer
ResNet-200 (Fast AA)	80.6	Fast AutoAugment
ResNet-200 (UA)	80.4	UniformAugment: A Search-free Probabilistic Data Augmentation Approach
ResNet-200 (AA)	80.0	AutoAugment: Learning Augmentation Policies from Data
ResNet-50 (DeepAA)	78.30	Deep AutoAugment
ResNet-50 (TA wide)	78.07	TrivialAugment: Tuning-free Yet State-of-the-Art Data Augmentation
ResNet-50 (LoRot-E)	77.72	Tailoring Self-Supervision for Supervised Learning
ResNet-50 (LoRot-I)	77.71	Tailoring Self-Supervision for Supervised Learning
ResNet-50 (UA)	77.63	UniformAugment: A Search-free Probabilistic Data Augmentation Approach
ResNet-50 (RA)	77.6	RandAugment: Practical automated data augmentation with a reduced search space
ResNet-50 (AA)	77.6	AutoAugment: Learning Augmentation Policies from Data
ResNet-50 (Fast AA)	77.6	Fast AutoAugment
ResNet-50 (DADA)	77.5	DADA: Differentiable Automatic Data Augmentation
ResNet-50 (Faster AA)	76.5	Faster AutoAugment: Learning Augmentation Strategies using Backpropagation
DeiT-T (+MixPro)	73.8	MixPro: Data Augmentation with MaskMix and Progressive Attention Labeling for Vision Transformer

0 of 17 row(s) selected.