# Image Classification on SVHN (test) (Few-Shot Evaluation)

Source: [Image Classification on SVHN (test) (Few-Shot Evaluation)](<https://www.sota2.com/research/sota/image-classification-on-svhn-test-few-shot-evaluation>)

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

Data updated: 2026-02-26T05:26:48.194Z

Evaluation of Image Classification performance on the test split of the SVHN dataset using few-shot learning settings (1, 5, 10, and 25 shots).

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) | SVHN (test) |
| Metric | Acc (1-shot) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [Acc (1-shot)](<https://www.sota2.com/research/sota/image-classification-on-svhn-test-few-shot-evaluation.md?metric=Acc%20(1-shot)>): Higher is better; unit: PERCENT.

- [Acc (5-shot)](<https://www.sota2.com/research/sota/image-classification-on-svhn-test-few-shot-evaluation.md?metric=Acc%20(5-shot)>): Higher is better; unit: PERCENT.

- [Acc (10-shot)](<https://www.sota2.com/research/sota/image-classification-on-svhn-test-few-shot-evaluation.md?metric=Acc%20(10-shot)>): Higher is better; unit: PERCENT.

- [Acc (25-shot)](<https://www.sota2.com/research/sota/image-classification-on-svhn-test-few-shot-evaluation.md?metric=Acc%20(25-shot)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Acc (1-shot) | Result date | Report paper |
| --- | --- | --- | --- |
| MAP^wd + FroFA | 21.8 | Not recorded | 68fa74411e209131cea608c8 |
| MAP^wd | 20.7 | Not recorded | 68fa74411e209131cea608c8 |
| Linear probe | 11.8 | Not recorded | 68fa74411e209131cea608c8 |

### MAP^wd + FroFA

Acc (1-shot): 21.8. Result ID: 699344023a6c31484914cf1a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74411e209131cea608c8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74411e209131cea608c8.md>)

[Implementation](<https://frozen-feature-augmentation.github.io/>)

Recorded evaluation configuration: {"Backbone":"JFT-3B L/16","training\_mode":"weight-decayed MAP with brightness c2FroFA"}. This record does not assert independent reproduction.

### MAP^wd

Acc (1-shot): 20.7. Result ID: 699344023a6c31484914cf18.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74411e209131cea608c8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65c31e209131cea0c4e7.md>)

Recorded evaluation configuration: {"Backbone":"JFT-3B L/16","training\_mode":"weight-decayed MAP"}. This record does not assert independent reproduction.

### Linear probe

Acc (1-shot): 11.8. Result ID: 699344023a6c31484914cf19.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74411e209131cea608c8.md>)

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

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

Recorded evaluation configuration: {"Backbone":"JFT-3B L/16","training\_mode":"L2-regularized linear probe"}. This record does not assert independent reproduction.

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