# Image Classification on STL10 (Clean, AutoAttack, PGD Robustness)

Source: [Image Classification on STL10 (Clean, AutoAttack, PGD Robustness)](<https://www.sota2.com/research/sota/image-classification-on-stl10-clean-autoattack-pgd-robustness>)

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

Data updated: 2026-02-27T11:46:09.288Z

Evaluation of Image Classification performance on the STL-10 dataset, reporting clean accuracy and robustness against AutoAttack and PGD adversarial attacks.

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.

| Scope | Recorded value |
| --- | --- |
| Task | Image Classification |
| Dataset (including any stated split) | STL10 |
| Metric | Clean Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Clean Accuracy](<https://www.sota2.com/research/sota/image-classification-on-stl10-clean-autoattack-pgd-robustness.md?metric=Clean%20Accuracy>): Higher is better; unit: PERCENT.

- [AutoAttack Robustness](<https://www.sota2.com/research/sota/image-classification-on-stl10-clean-autoattack-pgd-robustness.md?metric=AutoAttack%20Robustness>): Higher is better; unit: PERCENT.

- [PGD Robustness](<https://www.sota2.com/research/sota/image-classification-on-stl10-clean-autoattack-pgd-robustness.md?metric=PGD%20Robustness>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Clean Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CLIP | 97.03 | Not recorded | 6940e7a8456146936284f897 |
| FARE | 95.69 | Not recorded | 6940e7a8456146936284f897 |
| UCAT | 95.65 | Not recorded | 6940e7a8456146936284f897 |
| TeCoA | 93.3 | Not recorded | 6940e7a8456146936284f897 |

### CLIP

Clean Accuracy: 97.03. Result ID: 69a183f135dfcabb00e788a5.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7a8456146936284f897.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Training Data":"ImageNet-1k","Adversarial Training":"2-step PGD","Epsilon":"1/255"}. This record does not assert independent reproduction.

### FARE

Clean Accuracy: 95.69. Result ID: 69a183f135dfcabb00e788a7.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7a8456146936284f897.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738c1e209131cea5d24a.md>)

Recorded evaluation configuration: {"Training Data":"ImageNet-1k","Adversarial Training":"2-step PGD","Epsilon":"1/255"}. This record does not assert independent reproduction.

### UCAT

Clean Accuracy: 95.65. Result ID: 69a183f135dfcabb00e788a8.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7a8456146936284f897.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6940e7a8456146936284f897.md>)

Recorded evaluation configuration: {"Training Data":"ImageNet-1k","Adversarial Training":"2-step PGD","Epsilon":"1/255"}. This record does not assert independent reproduction.

### TeCoA

Clean Accuracy: 93.3. Result ID: 69a183f135dfcabb00e788a6.

[Reporting paper](<https://www.sota2.com/research/paper/6940e7a8456146936284f897.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f271e209131cea44177.md>)

Recorded evaluation configuration: {"Training Data":"ImageNet-1k","Adversarial Training":"2-step PGD","Epsilon":"1/255"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
