Image Classification On Objectnet

Metrics

Top-1 Accuracy

Results

Performance results of various models on this benchmark

Model Name	Top-1 Accuracy	Paper Title	Repository
ResNet-50 + MixUp (rescaled)	28.37	On Mixup Regularization
MoCo-v2 (BG_Swaps)	20.8	Characterizing and Improving the Robustness of Self-Supervised Learning through Background Augmentations	-
AR-B (Opt Relevance)	47.1	Optimizing Relevance Maps of Vision Transformers Improves Robustness
RegViT (RandAug)	29.3	Pyramid Adversarial Training Improves ViT Performance
Vit B/16 (Bamboo)	53.9	Bamboo: Building Mega-Scale Vision Dataset Continually with Human-Machine Synergy
CLIP (CC12M pretrain)	15.24	Robust Cross-Modal Representation Learning with Progressive Self-Distillation	-
MLP-Mixer + Pixel	24.75	Pyramid Adversarial Training Improves ViT Performance
ALIGN	72.2	Combined Scaling for Zero-shot Transfer Learning	-
RegNetY 128GF (Platt)	64.3	Revisiting Weakly Supervised Pre-Training of Visual Perception Models
ViT H/14 (Platt)	60	Revisiting Weakly Supervised Pre-Training of Visual Perception Models
NASNet-A	35.77	ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models	-
SWAG (ViT H/14)	69.5	Revisiting Weakly Supervised Pre-Training of Visual Perception Models
SwAV (reverse linear probing)	17.71	Measuring the Interpretability of Unsupervised Representations via Quantized Reversed Probing	-
BYOL (BG_RM)	23.9	Characterizing and Improving the Robustness of Self-Supervised Learning through Background Augmentations	-
Inception-v4	32.24	ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models	-
AlexNet	6.78	ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models	-
Discrete ViT	29.95	Pyramid Adversarial Training Improves ViT Performance
SwAV (BG_RM)	21.9	Characterizing and Improving the Robustness of Self-Supervised Learning through Background Augmentations	-
MAWS (ViT-H)	72.6	The effectiveness of MAE pre-pretraining for billion-scale pretraining
OBoW (reverse linear probing)	12.23	Measuring the Interpretability of Unsupervised Representations via Quantized Reversed Probing	-

0 of 106 row(s) selected.