# Image Classification on CIFAR-10.02

Source: [Image Classification on CIFAR-10.02](<https://www.sota2.com/research/sota/image-classification-on-cifar-10-02>)

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

Data updated: 2026-02-26T06:08:39.278Z

Evaluation of Image Classification performance on the CIFAR-10.02 dataset, reporting WG, Average, and Gap metrics.

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) | CIFAR-10.02 |
| Metric | WG Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [WG Score](<https://www.sota2.com/research/sota/image-classification-on-cifar-10-02.md?metric=WG%20Score>): Higher is better; unit: PERCENT.

- [Average Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-10-02.md?metric=Average%20Accuracy>): Higher is better; unit: PERCENT.

- [Performance Gap](<https://www.sota2.com/research/sota/image-classification-on-cifar-10-02.md?metric=Performance%20Gap>): Lower is better; unit: PERCENT.

## Recorded results

| Method | WG Score | Result date | Report paper |
| --- | --- | --- | --- |
| ERM Adapter | 68.8 | Not recorded | 68fa6d041e209131cea375ed |
| Contrastive Adapter | 60.7 | Not recorded | 68fa6d041e209131cea375ed |
| WiSE-FT | 58.2 | Not recorded | 68fa6d041e209131cea375ed |
| ERM LP | 51.3 | Not recorded | 68fa6d041e209131cea375ed |
| DFR (Sub) | 45 | Not recorded | 68fa6d041e209131cea375ed |
| DFR (Up) | 38.5 | Not recorded | 68fa6d041e209131cea375ed |
| Zero-Shot | 31.9 | Not recorded | 68fa6d041e209131cea375ed |

### ERM Adapter

WG Score: 68.8. Result ID: 698f790c3a6c314849f8e055.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77921e209131cea7403b.md>)

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Adapter"}. This record does not assert independent reproduction.

### Contrastive Adapter

WG Score: 60.7. Result ID: 698f790c3a6c314849f8e059.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Contrastive Adaptation"}. This record does not assert independent reproduction.

### WiSE-FT

WG Score: 58.2. Result ID: 698f790c3a6c314849f8e056.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd31e209131cea3601a.md>)

[Implementation](<https://github.com/mlfoundations/wise-ft>)

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Fine-tuning"}. This record does not assert independent reproduction.

### ERM LP

WG Score: 51.3. Result ID: 698f790c3a6c314849f8e054.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

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

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

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Linear Probing"}. This record does not assert independent reproduction.

### DFR (Sub)

WG Score: 45. Result ID: 698f790c3a6c314849f8e057.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fe41e209131cea47f22.md>)

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Subsample DFR"}. This record does not assert independent reproduction.

### DFR (Up)

WG Score: 38.5. Result ID: 698f790c3a6c314849f8e058.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fe41e209131cea47f22.md>)

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Upsample DFR"}. This record does not assert independent reproduction.

### Zero-Shot

WG Score: 31.9. Result ID: 698f790c3a6c314849f8e053.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d041e209131cea375ed.md>)

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

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

Recorded evaluation configuration: {"Backbone":"CLIP RN-50","Evaluation Protocol":"Zero-Shot"}. This record does not assert independent reproduction.

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