# Image Classification on ImageNet FGSM

Source: [Image Classification on ImageNet FGSM](<https://www.sota2.com/research/sota/image-classification-on-imagenet-fgsm>)

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

Data updated: 2026-04-21T08:33:52.687Z

Evaluation of Image Classification performance on the ImageNet FGSM dataset, reporting Accuracy.

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) | ImageNet FGSM |
| Metric | Top-1 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [Top-1 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-imagenet-fgsm.md?metric=Top-1%20Accuracy>): Higher is better; unit: PERCENT.

- [Accuracy](<https://www.sota2.com/research/sota/image-classification-on-imagenet-fgsm.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Top-5 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-imagenet-fgsm.md?metric=Top-5%20Accuracy>): Higher is better; unit: PERCENT.

- [NLL](<https://www.sota2.com/research/sota/image-classification-on-imagenet-fgsm.md?metric=NLL>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Top-1 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| QUEST | 56.9 | Not recorded | 69cdf7f20636e5bf4afcc99a |
| Elliptical-QUEST | 56.39 | Not recorded | 69cdf7f20636e5bf4afcc99a |
| Elliptical | 55.96 | Not recorded | 69cdf7f20636e5bf4afcc99a |
| Swin-ACMoE-Top 2 | 55.78 | Not recorded | 68fa77411e209131cea72040 |
| Swin-Top 2 | 54.7 | Not recorded | 68fa77411e209131cea72040 |
| Standard | 54.23 | Not recorded | 69cdf7f20636e5bf4afcc99a |
| Swin-ACMoE-Top 1 | 53.43 | Not recorded | 68fa77411e209131cea72040 |
| Swin-Top 1 | 52.84 | Not recorded | 68fa77411e209131cea72040 |

### QUEST

Top-1 Accuracy: 56.9. Result ID: 69ce2bef47f34ce924b75d2b.

[Reporting paper](<https://www.sota2.com/research/paper/69cdf7f20636e5bf4afcc99a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cdf7f20636e5bf4afcc99a.md>)

Recorded evaluation configuration: {"Backbone":"ViT-Ti/16","Training Recipe":"DeiT","Attention mechanism":"QUEST"}. This record does not assert independent reproduction.

### Elliptical-QUEST

Top-1 Accuracy: 56.39. Result ID: 69ce2bef47f34ce924b75d2d.

[Reporting paper](<https://www.sota2.com/research/paper/69cdf7f20636e5bf4afcc99a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cdf7f20636e5bf4afcc99a.md>)

Recorded evaluation configuration: {"Backbone":"ViT-Ti/16","Training Recipe":"DeiT","Attention mechanism":"Elliptical-QUEST"}. This record does not assert independent reproduction.

### Elliptical

Top-1 Accuracy: 55.96. Result ID: 69ce2bef47f34ce924b75d2c.

[Reporting paper](<https://www.sota2.com/research/paper/69cdf7f20636e5bf4afcc99a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75751e209131cea680fa.md>)

Recorded evaluation configuration: {"Backbone":"ViT-Ti/16","Training Recipe":"DeiT","Attention mechanism":"Elliptical"}. This record does not assert independent reproduction.

### Swin-ACMoE-Top 2

Top-1 Accuracy: 55.78. Result ID: 69e7366a47f34ce924cd6ead.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72040.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72040.md>)

Recorded evaluation configuration: {"Top-k routing":"2"}. This record does not assert independent reproduction.

### Swin-Top 2

Top-1 Accuracy: 54.7. Result ID: 69e7366a47f34ce924cd6eac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72040.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a8c1e209131cea28b62.md>)

[Implementation](<https://github.com/microsoft/Swin-Transformer>)

Recorded evaluation configuration: {"Top-k routing":"2"}. This record does not assert independent reproduction.

### Standard

Top-1 Accuracy: 54.23. Result ID: 69ce2bef47f34ce924b75d2a.

[Reporting paper](<https://www.sota2.com/research/paper/69cdf7f20636e5bf4afcc99a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba01e209131cea8967c.md>)

Recorded evaluation configuration: {"Backbone":"ViT-Ti/16","Training Recipe":"DeiT","Attention mechanism":"Standard"}. This record does not assert independent reproduction.

### Swin-ACMoE-Top 1

Top-1 Accuracy: 53.43. Result ID: 69e7366a47f34ce924cd6eab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72040.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72040.md>)

Recorded evaluation configuration: {"Top-k routing":"1"}. This record does not assert independent reproduction.

### Swin-Top 1

Top-1 Accuracy: 52.84. Result ID: 69e7366a47f34ce924cd6eaa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72040.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a8c1e209131cea28b62.md>)

[Implementation](<https://github.com/microsoft/Swin-Transformer>)

Recorded evaluation configuration: {"Top-k routing":"1"}. This record does not assert independent reproduction.

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