# Image Classification on OOD Benchmark Suite Subset (Cars, DTD, EuroSAT, GTSRB, MNIST, RESISC45, SVHN)

Source: [Image Classification on OOD Benchmark Suite Subset (Cars, DTD, EuroSAT, GTSRB, MNIST, RESISC45, SVHN)](<https://www.sota2.com/research/sota/image-classification-on-ood-benchmark-suite-subset-cars-dtd-eurosat-gtsrb-mnist-resisc45-svhn>)

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

Data updated: 2026-04-15T06:52:07.006Z

Aggregate performance of Image Classification models evaluated across a specific subset of seven datasets (Cars, DTD, EuroSAT, GTSRB, MNIST, RESISC45, SVHN), typically used for Out-of-Distribution (OOD) generalization testing, measured by Average OOD 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) | OOD benchmark suite (SVHN, Cars, RESISC45, EuroSAT, SUN397, GTSRB, MNIST, DTD) |
| Metric | Average OOD Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 16 of 16 |
| 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

- [Average OOD Accuracy](<https://www.sota2.com/research/sota/image-classification-on-ood-benchmark-suite-subset-cars-dtd-eurosat-gtsrb-mnist-resisc45-svhn.md?metric=Average%20OOD%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Average OOD Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| FlexMerge | 74.88 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 74.64 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 74.53 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 74.16 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 74.03 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 73.98 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 73.94 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 73.9 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 65.95 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 65.86 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 65.67 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 65.48 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 65.39 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 64.75 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 64.1 | Not recorded | 68fa78ee1e209131cea7b48d |
| FlexMerge | 62.89 | Not recorded | 68fa78ee1e209131cea7b48d |

### FlexMerge

Average OOD Accuracy: 74.88. Result ID: 69df33e147f34ce924c77ab9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"8x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 74.64. Result ID: 69df33e147f34ce924c77ab8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"6x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 74.53. Result ID: 69df33e147f34ce924c77ab7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"4x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 74.16. Result ID: 69df33e147f34ce924c77ab6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"3x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 74.03. Result ID: 69df33e147f34ce924c77ab5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"2.5x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 73.98. Result ID: 69df33e147f34ce924c77ab4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"2x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 73.94. Result ID: 69df33e147f34ce924c77ab3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"1.5x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 73.9. Result ID: 69df33e147f34ce924c77ab2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-L-14","Merged Size":"1x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 65.95. Result ID: 69df33e147f34ce924c77ab1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"8x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 65.86. Result ID: 69df33e147f34ce924c77aaf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"4x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 65.67. Result ID: 69df33e147f34ce924c77aae.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"3x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 65.48. Result ID: 69df33e147f34ce924c77ab0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"6x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 65.39. Result ID: 69df33e147f34ce924c77aad.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"2.5x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 64.75. Result ID: 69df33e147f34ce924c77aac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"2x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 64.1. Result ID: 69df33e147f34ce924c77aab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"1.5x"}. This record does not assert independent reproduction.

### FlexMerge

Average OOD Accuracy: 62.89. Result ID: 69df33e147f34ce924c77aaa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78ee1e209131cea7b48d.md>)

Recorded evaluation configuration: {"Model":"ViT-B-32","Merged Size":"1x"}. This record does not assert independent reproduction.

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