# How good is my GAN?

Source: [How good is my GAN?](<https://www.sota2.com/research/paper/68fa649b1e209131cea057fe>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T21:08:02.971Z

Authors: Konstantin Shmelkov, Cordelia Schmid, Karteek Alahari

Published: 2018-07-25T09:31:17.000Z

## Abstract

Generative adversarial networks (GANs) are one of the most popular methods for generating images today. While impressive results have been validated by visual inspection, a number of quantitative criteria have emerged only recently. We argue here that the existing ones are insufficient and need to be in adequation with the task at hand. In this paper we introduce two measures based on image classification---GAN-train and GAN-test, which approximate the recall (diversity) and precision (quality of the image) of GANs respectively. We evaluate a number of recent GAN approaches based on these two measures and demonstrate a clear difference in performance. Furthermore, we observe that the increasing difficulty of the dataset, from CIFAR10 over CIFAR100 to ImageNet, shows an inverse correlation with the quality of the GANs, as clearly evident from our measures.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Image Classification on Stanford Cars 10% subset](<https://www.sota2.com/research/sota/image-classification-on-stanford-cars-10-percent-subset.md>): Image Classification; Stanford Cars 10% subset; Top-1 Acc: 18.91.

- [Classification on CIFAR-10-C 1.0 (test)](<https://www.sota2.com/research/sota/classification-on-cifar-10-c-1-0-test.md>): Classification; CIFAR-10-C 1.0 (test); Accuracy (Clean): 84.11.

- [Image Classification on Stanford Cars (25% subset)](<https://www.sota2.com/research/sota/image-classification-on-stanford-cars-25-percent-subset.md>): Image Classification; Stanford Cars 25% subset; Top-1 Accuracy: 46.56.

- [Image Classification on Stanford Cars 50%](<https://www.sota2.com/research/sota/image-classification-on-stanford-cars-50-percent.md>): Image Classification; Stanford Cars 50%; Top-1 Accuracy: 68.81.

- [Image Classification on Stanford Cars 75%](<https://www.sota2.com/research/sota/image-classification-on-stanford-cars-75-percent.md>): Image Classification; Stanford Cars 75%; Top-1 Accuracy: 80.67.

- [Image Classification on ImageNet (val) (ACC)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-val-acc.md>): Image Classification; ImageNet (val); Accuracy: not recorded.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1807.09499>)

[PDF](<https://arxiv.org/pdf/1807.09499.pdf>)