# Image Classification on CIFAR-10 (test) (Clean and PGD Robustness)

Source: [Image Classification on CIFAR-10 (test) (Clean and PGD Robustness)](<https://www.sota2.com/research/sota/image-classification-on-cifar-10-test-clean-and-pgd-robustness>)

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

Data updated: 2026-06-16T06:59:04.202Z

Evaluation of Image Classification performance on the CIFAR-10 test set, reporting standard Clean Accuracy and robustness against PGD attacks (PGD-20).

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) | CIFAR-10 (test) |
| Metric | Clean Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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-cifar-10-test-clean-and-pgd-robustness.md?metric=Clean%20Accuracy>): Higher is better; unit: PERCENT.

- [PGD-20 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-10-test-clean-and-pgd-robustness.md?metric=PGD-20%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Clean Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| AT-Grow | 82.82 | Not recorded | 6a30d893d33dfb23e783cc25 |
| Fixed ResNet-18 AT | 82.74 | Not recorded | 6a30d893d33dfb23e783cc25 |
| Stabilized Seq-Grow | 82.51 | Not recorded | 6a30d893d33dfb23e783cc25 |
| GRAPE | 82.43 | Not recorded | 6a30d893d33dfb23e783cc25 |
| Stabilized AT | 81.71 | Not recorded | 6a30d893d33dfb23e783cc25 |

### AT-Grow

Clean Accuracy: 82.82. Result ID: 6a30f32ffbc373592c2c932e.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d893d33dfb23e783cc25.md>)

Recorded evaluation configuration: {"FLOPs Ratio":"0.996×","Params Ratio":"0.748×"}. This record does not assert independent reproduction.

### Fixed ResNet-18 AT

Clean Accuracy: 82.74. Result ID: 6a30f32ffbc373592c2c932d.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d893d33dfb23e783cc25.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665a1e209131cea0fbae.md>)

Recorded evaluation configuration: {"FLOPs Ratio":"1.000×","Params Ratio":"1.000×"}. This record does not assert independent reproduction.

### Stabilized Seq-Grow

Clean Accuracy: 82.51. Result ID: 6a30f32ffbc373592c2c9330.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d893d33dfb23e783cc25.md>)

Recorded evaluation configuration: {"FLOPs Ratio":"1.000×","Params Ratio":"1.000×"}. This record does not assert independent reproduction.

### GRAPE

Clean Accuracy: 82.43. Result ID: 6a30f32ffbc373592c2c9331.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d893d33dfb23e783cc25.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a30d893d33dfb23e783cc25.md>)

Recorded evaluation configuration: {"FLOPs Ratio":"1.009×","Params Ratio":"0.786×"}. This record does not assert independent reproduction.

### Stabilized AT

Clean Accuracy: 81.71. Result ID: 6a30f32ffbc373592c2c932f.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d893d33dfb23e783cc25.md>)

Recorded evaluation configuration: {"FLOPs Ratio":"1.000×","Params Ratio":"1.000×"}. This record does not assert independent reproduction.

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