# Image Classification on ImageNet 2012 (val) (Policy Comparison)

Source: [Image Classification on ImageNet 2012 (val) (Policy Comparison)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison>)

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

Data updated: 2026-02-26T06:07:20.322Z

Evaluation of Image Classification performance on the validation split of ImageNet 2012, specifically reporting baseline errors, policy-adjusted errors, and the resulting gains.

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 2012 (val) |
| Metric | Top-1 Policy Error |
| Unit | PERCENT |
| Direction | Lower 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

- [Top-1 Policy Error](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison.md?metric=Top-1%20Policy%20Error>): Lower is better; unit: PERCENT.

- [Top-1 Baseline Error](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison.md?metric=Top-1%20Baseline%20Error>): Lower is better; unit: PERCENT.

- [Top-5 Baseline Error](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison.md?metric=Top-5%20Baseline%20Error>): Lower is better; unit: PERCENT.

- [Top-5 Policy Error](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison.md?metric=Top-5%20Policy%20Error>): Lower is better; unit: PERCENT.

- [Top-1 Baseline Improvement](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison.md?metric=Top-1%20Baseline%20Improvement>): Higher is better; unit: PERCENT.

- [Top-5 Baseline Improvement](<https://www.sota2.com/research/sota/image-classification-on-imagenet-2012-val-policy-comparison.md?metric=Top-5%20Baseline%20Improvement>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Top-1 Policy Error | Result date | Report paper |
| --- | --- | --- | --- |
| AA | 22.4 | Not recorded | 68fa66aa1e209131cea117a8 |
| Fast AA | 22.4 | Not recorded | 68fa66aa1e209131cea117a8 |
| Faster AutoAugment | 23.5 | Not recorded | 68fa66aa1e209131cea117a8 |

### AA

Top-1 Policy Error: 22.4. Result ID: 698cfe198b517eb4331e3777.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66aa1e209131cea117a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65a51e209131cea0b94a.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction.

### Fast AA

Top-1 Policy Error: 22.4. Result ID: 698cfe198b517eb4331e3778.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66aa1e209131cea117a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66e21e209131cea12d0c.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction.

### Faster AutoAugment

Top-1 Policy Error: 23.5. Result ID: 698cfe198b517eb4331e3779.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66aa1e209131cea117a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66aa1e209131cea117a8.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Training Epochs":"200"}. This record does not assert independent reproduction.

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