# Robustness Evaluation on ImageNet Robustness Variants (IN-A, IN-R, IN-Sketch, IN-C) (test)

Source: [Robustness Evaluation on ImageNet Robustness Variants (IN-A, IN-R, IN-Sketch, IN-C) (test)](<https://www.sota2.com/research/sota/robustness-evaluation-on-imagenet-robustness-variants-in-a-in-r-in-sketch-in-c-test>)

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

Data updated: 2026-02-26T05:46:52.623Z

Simultaneous evaluation of model robustness across four key ImageNet variants: ImageNet-A (adversarial), ImageNet-R (renderings), ImageNet-Sketch, and ImageNet-C (corruptions), measured on the test split.

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 | Robustness Evaluation |
| Dataset (including any stated split) | ImageNet Robustness Variants (IN-A, IN-R, IN-Sketch, IN-C) (test) |
| Metric | IN-A Top-1 Acc |
| 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

- [IN-A Top-1 Acc](<https://www.sota2.com/research/sota/robustness-evaluation-on-imagenet-robustness-variants-in-a-in-r-in-sketch-in-c-test.md?metric=IN-A%20Top-1%20Acc>): Higher is better; unit: PERCENT.

- [IN-R Top-1 Acc](<https://www.sota2.com/research/sota/robustness-evaluation-on-imagenet-robustness-variants-in-a-in-r-in-sketch-in-c-test.md?metric=IN-R%20Top-1%20Acc>): Higher is better; unit: PERCENT.

- [IN-Sketch Top-1 Acc](<https://www.sota2.com/research/sota/robustness-evaluation-on-imagenet-robustness-variants-in-a-in-r-in-sketch-in-c-test.md?metric=IN-Sketch%20Top-1%20Acc>): Higher is better; unit: PERCENT.

- [IN-C mCE](<https://www.sota2.com/research/sota/robustness-evaluation-on-imagenet-robustness-variants-in-a-in-r-in-sketch-in-c-test.md?metric=IN-C%20mCE>): Lower is better; unit: PERCENT.

## Recorded results

| Method | IN-A Top-1 Acc | Result date | Report paper |
| --- | --- | --- | --- |
| SiameseIM | 43.8 | Not recorded | 68fa6e031e209131cea3d696 |
| SiameseIM | 38.6 | Not recorded | 68fa6e031e209131cea3d696 |
| MSN | 37.5 | Not recorded | 68fa6e031e209131cea3d696 |
| MAE | 35.9 | Not recorded | 68fa6e031e209131cea3d696 |
| MoCo-v3 | 32.4 | Not recorded | 68fa6e031e209131cea3d696 |

### SiameseIM

IN-A Top-1 Acc: 43.8. Result ID: 698ff4713a6c314849fd0cf3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Implementation](<https://github.com/fundamentalvision/Siamese-Image-Modeling>)

Recorded evaluation configuration: {"Epochs":"1600"}. This record does not assert independent reproduction.

### SiameseIM

IN-A Top-1 Acc: 38.6. Result ID: 698ff4713a6c314849fd0cf2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Implementation](<https://github.com/fundamentalvision/Siamese-Image-Modeling>)

Recorded evaluation configuration: {"Epochs":"400"}. This record does not assert independent reproduction.

### MSN

IN-A Top-1 Acc: 37.5. Result ID: 698ff4713a6c314849fd0cef.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673f1e209131cea15344.md>)

[Implementation](<https://github.com/facebookresearch/moco>)

Recorded evaluation configuration: {"Epochs":"1200","multi-crop":true,"symmetric loss":true}. This record does not assert independent reproduction.

### MAE

IN-A Top-1 Acc: 35.9. Result ID: 698ff4713a6c314849fd0cf1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b7f1e209131cea2e973.md>)

Recorded evaluation configuration: {"Epochs":"1600"}. This record does not assert independent reproduction.

### MoCo-v3

IN-A Top-1 Acc: 32.4. Result ID: 698ff4713a6c314849fd0cf0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e031e209131cea3d696.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c61e209131cea1e54d.md>)

Recorded evaluation configuration: {"Epochs":"600","symmetric loss":true}. This record does not assert independent reproduction.

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