TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
Image Classification	CIFAR-10	DCGAN	Percentage correct	82.8	# 210
Conditional Image Generation	CIFAR-10	DCGAN	Inception score	6.58	# 18
Image Clustering	CIFAR-10	GAN	Accuracy	0.315	# 26
Image Clustering	CIFAR-10	GAN	NMI	0.265	# 22
Image Clustering	CIFAR-10	GAN	Train set	Train+Test	# 1
Image Clustering	CIFAR-10	GAN	ARI	0.176	# 22
Image Clustering	CIFAR-10	GAN	Backbone	GAN	# 1
Image Classification	CIFAR-10	1 Layer K-means	Percentage correct	80.6	# 215
Image Clustering	CIFAR-100	GAN	Accuracy	0.151	# 22
Image Clustering	CIFAR-100	GAN	NMI	0.120	# 18
Image Clustering	CIFAR-100	GAN	Train Set	Train+Test	# 1
Image Clustering	ImageNet-10	GAN	Accuracy	0.346	# 14
Image Clustering	ImageNet-10	GAN	NMI	0.225	# 14
Image Clustering	Imagenet-dog-15	GAN	Accuracy	0.174	# 17
Image Clustering	Imagenet-dog-15	GAN	NMI	0.121	# 16
Image Clustering	STL-10	GAN	Accuracy	0.298	# 22
Image Clustering	STL-10	GAN	NMI	0.210	# 19
Image Clustering	STL-10	GAN	Train Split	Train+Test	# 1
Image Classification	SVHN	TSVM	Percentage error	66.55	# 57
Image Classification	SVHN	DCGAN	Percentage error	22.48	# 50
Image Classification	SVHN	KNN	Percentage error	77.93	# 58
Image Classification	SVHN	Supervised CNN	Percentage error	28.87	# 52
Image Clustering	Tiny-ImageNet	GAN	Accuracy	0.041	# 9
Image Clustering	Tiny-ImageNet	GAN	NMI	0.135	# 9

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-clustering-on-tiny-imagenet)](https://paperswithcode.com/sota/image-clustering-on-tiny-imagenet?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-clustering-on-imagenet-10)](https://paperswithcode.com/sota/image-clustering-on-imagenet-10?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-clustering-on-imagenet-dog-15)](https://paperswithcode.com/sota/image-clustering-on-imagenet-dog-15?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/conditional-image-generation-on-cifar-10)](https://paperswithcode.com/sota/conditional-image-generation-on-cifar-10?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-clustering-on-cifar-100)](https://paperswithcode.com/sota/image-clustering-on-cifar-100?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-clustering-on-stl-10)](https://paperswithcode.com/sota/image-clustering-on-stl-10?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-clustering-on-cifar-10)](https://paperswithcode.com/sota/image-clustering-on-cifar-10?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-classification-on-svhn)](https://paperswithcode.com/sota/image-classification-on-svhn?p=unsupervised-representation-learning-with-1)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-representation-learning-with-1/image-classification-on-cifar-10)](https://paperswithcode.com/sota/image-classification-on-cifar-10?p=unsupervised-representation-learning-with-1)`

Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

19 Nov 2015 · Alec Radford, Luke Metz, Soumith Chintala ·

In recent years, supervised learning with convolutional networks (CNNs) has seen huge adoption in computer vision applications. Comparatively, unsupervised learning with CNNs has received less attention. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised learning. We introduce a class of CNNs called deep convolutional generative adversarial networks (DCGANs), that have certain architectural constraints, and demonstrate that they are a strong candidate for unsupervised learning. Training on various image datasets, we show convincing evidence that our deep convolutional adversarial pair learns a hierarchy of representations from object parts to scenes in both the generator and discriminator. Additionally, we use the learned features for novel tasks - demonstrating their applicability as general image representations.

PDF Abstract

Code

Add Remove Mark official

tensorflow/models

76,582

labmlai/annotated_deep_learning_pap…

↳ View annotated code at

labml.ai

47,627

eriklindernoren/PyTorch-GAN

15,676

eriklindernoren/Keras-GAN

9,135

carpedm20/DCGAN-tensorflow

7,130

See all 261 implementations

Tasks

Add Remove

Conditional Image Generation

Image Clustering

Representation Learning

Datasets

CIFAR-10

ImageNet

CIFAR-100

SVHN

STL-10

Tiny ImageNet

Results from the Paper

Edit

Ranked #9 on Image Clustering on Tiny-ImageNet

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
Image Classification	CIFAR-10	DCGAN	Percentage correct	82.8	# 210	Compare
Conditional Image Generation	CIFAR-10	DCGAN	Inception score	6.58	# 18	Compare
Image Clustering	CIFAR-10	GAN	Accuracy	0.315	# 26	Compare
			NMI	0.265	# 22	Compare
			Train set	Train+Test	# 1	Compare
			ARI	0.176	# 22	Compare
			Backbone	GAN	# 1	Compare
Image Classification	CIFAR-10	1 Layer K-means	Percentage correct	80.6	# 215	Compare
Image Classification	SVHN	TSVM	Percentage error	66.55	# 57	Compare
Image Classification	SVHN	DCGAN	Percentage error	22.48	# 50	Compare
Image Classification	SVHN	KNN	Percentage error	77.93	# 58	Compare
Image Classification	SVHN	Supervised CNN	Percentage error	28.87	# 52	Compare

Results from Other Papers

Task	Dataset	Model	Metric Name	Metric Value	Rank	Compare
Image Clustering	CIFAR-100	GAN	Accuracy	0.151	# 22	See all
			NMI	0.120	# 18	See all
			Train Set	Train+Test	# 1	See all
Image Clustering	ImageNet-10	GAN	Accuracy	0.346	# 14	See all
Image Clustering	ImageNet-10	GAN	NMI	0.225	# 14	See all
Image Clustering	Imagenet-dog-15	GAN	Accuracy	0.174	# 17	See all
Image Clustering	Imagenet-dog-15	GAN	NMI	0.121	# 16	See all
Image Clustering	STL-10	GAN	Accuracy	0.298	# 22	See all
			NMI	0.210	# 19	See all
			Train Split	Train+Test	# 1	See all
Image Clustering	Tiny-ImageNet	GAN	Accuracy	0.041	# 9	See all
Image Clustering	Tiny-ImageNet	GAN	NMI	0.135	# 9	See all

Methods

Add Remove

Adam • Batch Normalization • Convolution • DCGAN • Leaky ReLU • ReLU • SVM • Weight Decay

Edit Social Preview

Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit