# Image Classification on SVHN (Natural Accuracy and Regularization Analysis)

Source: [Image Classification on SVHN (Natural Accuracy and Regularization Analysis)](<https://www.sota2.com/research/sota/image-classification-on-svhn-natural-accuracy-and-regularization-analysis>)

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

Data updated: 2026-02-26T05:48:03.482Z

Evaluation of standard Image Classification performance on SVHN, reporting natural accuracy alongside metrics derived from model regularization, such as L2 weight norm and specific penalized cross-entropy terms.

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 |
| Metric | L2 Weight Norm |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [L2 Weight Norm](<https://www.sota2.com/research/sota/image-classification-on-svhn-natural-accuracy-and-regularization-analysis.md?metric=L2%20Weight%20Norm>): Higher is better; unit: SCALAR.

- [Accuracy (Natural)](<https://www.sota2.com/research/sota/image-classification-on-svhn-natural-accuracy-and-regularization-analysis.md?metric=Accuracy%20(Natural)>): Higher is better; unit: PERCENT.

- [AutoAtt](<https://www.sota2.com/research/sota/image-classification-on-svhn-natural-accuracy-and-regularization-analysis.md?metric=AutoAtt>): Higher is better; unit: PERCENT.

- [Penalized Xent Loss](<https://www.sota2.com/research/sota/image-classification-on-svhn-natural-accuracy-and-regularization-analysis.md?metric=Penalized%20Xent%20Loss>): Lower is better; unit: SCALAR.

## Recorded results

| Method | L2 Weight Norm | Result date | Report paper |
| --- | --- | --- | --- |
| Non-adaptive Weight Decay | 102.11 | Not recorded | 68fa71551e209131cea508f2 |
| Adaptive Weight Decay (AWD) | 5.39 | Not recorded | 68fa71551e209131cea508f2 |

### Non-adaptive Weight Decay

L2 Weight Norm: 102.11. Result ID: 6991c5973a6c3148490a7e61.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71551e209131cea508f2.md>)

Recorded evaluation configuration: {"lambda\_wd":"5e-7","Optimizer":"SGD"}. This record does not assert independent reproduction.

### Adaptive Weight Decay (AWD)

L2 Weight Norm: 5.39. Result ID: 6991c5973a6c3148490a7e62.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71551e209131cea508f2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71551e209131cea508f2.md>)

Recorded evaluation configuration: {"lambda\_awd":"0.02378","Optimizer":"SGD"}. This record does not assert independent reproduction.

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