TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
Image Classification	CIFAR-10	TransBoost-ResNet50	Percentage correct	97.61	# 69
Image Classification	DTD	TransBoost-ResNet50	Accuracy	76.49	# 8
Image Classification	FGVC Aircraft	TransBoost-ResNet50	Accuracy	83.80%	# 1
Image Classification	Flowers-102	TransBoost-ResNet50	Accuracy	97.85%	# 33
Image Classification	Food-101	TransBoost-ResNet50	Accuracy (%)	84.30	# 4
Image Classification	ImageNet	TransBoost-Swin-T	Top 1 Accuracy	82.16%	# 522
Image Classification	ImageNet	TransBoost-Swin-T	Number of params	71.71M	# 790
Image Classification	ImageNet	TransBoost-ViT-S	Top 1 Accuracy	83.67%	# 374
Image Classification	ImageNet	TransBoost-ViT-S	Number of params	22.05M	# 563
Image Classification	ImageNet	TransBoost-ConvNext-T	Top 1 Accuracy	82.46%	# 490
Image Classification	ImageNet	TransBoost-ConvNext-T	Number of params	28.59M	# 640
Image Classification	ImageNet	TransBoost-ResNet152	Top 1 Accuracy	80.64%	# 633
Image Classification	ImageNet	TransBoost-ResNet152	Number of params	60.19M	# 767
Image Classification	ImageNet	TransBoost-MobileNetV3-L	Top 1 Accuracy	76.81%	# 827
Image Classification	ImageNet	TransBoost-MobileNetV3-L	Number of params	5.48M	# 422
Image Classification	ImageNet	TransBoost-ResNet34	Top 1 Accuracy	76.70%	# 832
Image Classification	ImageNet	TransBoost-ResNet34	Number of params	21.8M	# 553
Image Classification	ImageNet	TransBoost-ResNet18	Top 1 Accuracy	73.36%	# 914
Image Classification	ImageNet	TransBoost-ResNet18	Number of params	11.69M	# 491
Image Classification	ImageNet	TransBoost-EfficientNetB0	Top 1 Accuracy	78.60%	# 753
Image Classification	ImageNet	TransBoost-EfficientNetB0	Number of params	5.29M	# 414
Image Classification	ImageNet	TransBoost-ResNet101	Top 1 Accuracy	79.86%	# 674
Image Classification	ImageNet	TransBoost-ResNet101	Number of params	44.55M	# 701
Image Classification	ImageNet	TransBoost-ResNet50-StrikesBack	Top 1 Accuracy	81.15%	# 604
Image Classification	ImageNet	TransBoost-ResNet50-StrikesBack	Number of params	25.56M	# 597
Image Classification	ImageNet	TransBoost-ResNet50	Top 1 Accuracy	79.03%	# 724
Image Classification	Stanford Cars	TransBoost-ResNet50	Accuracy	90.80%	# 12
Image Classification	SUN397	TransBoost-ResNet50	Accuracy	95.94%	# 1

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-fgvc-aircraft)](https://paperswithcode.com/sota/image-classification-on-fgvc-aircraft?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-sun397)](https://paperswithcode.com/sota/image-classification-on-sun397?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-food-101-1)](https://paperswithcode.com/sota/image-classification-on-food-101-1?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-dtd)](https://paperswithcode.com/sota/image-classification-on-dtd?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-stanford-cars)](https://paperswithcode.com/sota/image-classification-on-stanford-cars?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-flowers-102)](https://paperswithcode.com/sota/image-classification-on-flowers-102?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-cifar-10)](https://paperswithcode.com/sota/image-classification-on-cifar-10?p=transboost-improving-the-best-imagenet)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/transboost-improving-the-best-imagenet/image-classification-on-imagenet)](https://paperswithcode.com/sota/image-classification-on-imagenet?p=transboost-improving-the-best-imagenet)`

TransBoost: Improving the Best ImageNet Performance using Deep Transduction

26 May 2022 · Omer Belhasin, Guy Bar-Shalom, Ran El-Yaniv ·

This paper deals with deep transductive learning, and proposes TransBoost as a procedure for fine-tuning any deep neural model to improve its performance on any (unlabeled) test set provided at training time. TransBoost is inspired by a large margin principle and is efficient and simple to use. Our method significantly improves the ImageNet classification performance on a wide range of architectures, such as ResNets, MobileNetV3-L, EfficientNetB0, ViT-S, and ConvNext-T, leading to state-of-the-art transductive performance. Additionally we show that TransBoost is effective on a wide variety of image classification datasets. The implementation of TransBoost is provided at: https://github.com/omerb01/TransBoost .

PDF Abstract

Code

Add Remove Mark official

omerb01/transboost official

Tasks

Add Remove

Image Classification

Transductive Learning

Datasets

CIFAR-10

ImageNet

CIFAR-100

Oxford 102 Flower

Stanford Cars

DTD

Food-101

FGVC-Aircraft

SUN397

Results from the Paper

Edit

Ranked #1 on Image Classification on SUN397

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
Image Classification	CIFAR-10	TransBoost-ResNet50	Percentage correct	97.61	# 69	Compare
Image Classification	DTD	TransBoost-ResNet50	Accuracy	76.49	# 8	Compare
Image Classification	FGVC Aircraft	TransBoost-ResNet50	Accuracy	83.80%	# 1	Compare
Image Classification	Flowers-102	TransBoost-ResNet50	Accuracy	97.85%	# 33	Compare
Image Classification	Food-101	TransBoost-ResNet50	Accuracy (%)	84.30	# 4	Compare
Image Classification	ImageNet	TransBoost-Swin-T	Top 1 Accuracy	82.16%	# 522	Compare
Image Classification	ImageNet	TransBoost-Swin-T	Number of params	71.71M	# 790	Compare
Image Classification	ImageNet	TransBoost-ViT-S	Top 1 Accuracy	83.67%	# 374	Compare
Image Classification	ImageNet	TransBoost-ViT-S	Number of params	22.05M	# 563	Compare
Image Classification	ImageNet	TransBoost-ConvNext-T	Top 1 Accuracy	82.46%	# 490	Compare
Image Classification	ImageNet	TransBoost-ConvNext-T	Number of params	28.59M	# 640	Compare
Image Classification	ImageNet	TransBoost-ResNet152	Top 1 Accuracy	80.64%	# 633	Compare
Image Classification	ImageNet	TransBoost-ResNet152	Number of params	60.19M	# 767	Compare
Image Classification	ImageNet	TransBoost-MobileNetV3-L	Top 1 Accuracy	76.81%	# 827	Compare
Image Classification	ImageNet	TransBoost-MobileNetV3-L	Number of params	5.48M	# 422	Compare
Image Classification	ImageNet	TransBoost-ResNet34	Top 1 Accuracy	76.70%	# 832	Compare
Image Classification	ImageNet	TransBoost-ResNet34	Number of params	21.8M	# 553	Compare
Image Classification	ImageNet	TransBoost-ResNet18	Top 1 Accuracy	73.36%	# 914	Compare
Image Classification	ImageNet	TransBoost-ResNet18	Number of params	11.69M	# 491	Compare
Image Classification	ImageNet	TransBoost-EfficientNetB0	Top 1 Accuracy	78.60%	# 753	Compare
Image Classification	ImageNet	TransBoost-EfficientNetB0	Number of params	5.29M	# 414	Compare
Image Classification	ImageNet	TransBoost-ResNet101	Top 1 Accuracy	79.86%	# 674	Compare
Image Classification	ImageNet	TransBoost-ResNet101	Number of params	44.55M	# 701	Compare
Image Classification	ImageNet	TransBoost-ResNet50-StrikesBack	Top 1 Accuracy	81.15%	# 604	Compare
Image Classification	ImageNet	TransBoost-ResNet50-StrikesBack	Number of params	25.56M	# 597	Compare
Image Classification	ImageNet	TransBoost-ResNet50	Top 1 Accuracy	79.03%	# 724	Compare
Image Classification	Stanford Cars	TransBoost-ResNet50	Accuracy	90.80%	# 12	Compare
Image Classification	SUN397	TransBoost-ResNet50	Accuracy	95.94%	# 1	Compare

Methods

Add Remove

No methods listed for this paper. Add relevant methods here

Edit Social Preview

TransBoost: Improving the Best ImageNet Performance using Deep Transduction

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit

Methods Edit Add Remove

Code

Add Remove Mark official

Tasks

Add Remove

Datasets

Results from the Paper

Edit

Methods

Add Remove