# Classification on CelebA HairColor (test)

Source: [Classification on CelebA HairColor (test)](<https://www.sota2.com/research/sota/classification-on-celeba-haircolor-test>)

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

Data updated: 2026-02-26T05:47:03.291Z

Evaluation of Classification performance on the test split of the CelebA HairColor dataset, reporting disparity metrics such as INDISTRIBUTION, WORST-GROUP, and GAP.

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 | Classification |
| Dataset (including any stated split) | CelebA HairColor (test) |
| Metric | In-Distribution Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [In-Distribution Accuracy](<https://www.sota2.com/research/sota/classification-on-celeba-haircolor-test.md?metric=In-Distribution%20Accuracy>): Higher is better; unit: PERCENT.

- [Worst-Group Accuracy](<https://www.sota2.com/research/sota/classification-on-celeba-haircolor-test.md?metric=Worst-Group%20Accuracy>): Higher is better; unit: PERCENT.

- [Accuracy Gap](<https://www.sota2.com/research/sota/classification-on-celeba-haircolor-test.md?metric=Accuracy%20Gap>): Lower is better; unit: SCALAR.

## Recorded results

| Method | In-Distribution Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| ERM | 95.6 | Not recorded | 68fa6f331e209131cea44667 |
| SSL+ERM | 95.5 | Not recorded | 68fa6f331e209131cea44667 |
| Group DRO | 93.1 | Not recorded | 68fa6f331e209131cea44667 |
| SSA | 92.8 | Not recorded | 68fa6f331e209131cea44667 |
| EIIL | 91.9 | Not recorded | 68fa6f331e209131cea44667 |
| LWBC | 88.9 | Not recorded | 68fa6f331e209131cea44667 |
| JTT | 88 | Not recorded | 68fa6f331e209131cea44667 |
| LfF | 86 | Not recorded | 68fa6f331e209131cea44667 |
| CVaR DRO | 82.4 | Not recorded | 68fa6f331e209131cea44667 |

### ERM

In-Distribution Accuracy: 95.6. Result ID: 6990afee3a6c3148490275ea.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":false}. This record does not assert independent reproduction.

### SSL+ERM

In-Distribution Accuracy: 95.5. Result ID: 6990afee3a6c3148490275ef.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet18","Spurious attribute label":false}. This record does not assert independent reproduction.

### Group DRO

In-Distribution Accuracy: 93.1. Result ID: 6990afee3a6c3148490275e8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":true}. This record does not assert independent reproduction.

### SSA

In-Distribution Accuracy: 92.8. Result ID: 6990afee3a6c3148490275e9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c2b1e209131cea3281e.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":true}. This record does not assert independent reproduction.

### EIIL

In-Distribution Accuracy: 91.9. Result ID: 6990afee3a6c3148490275ed.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a5f1e209131cea27cee.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":false}. This record does not assert independent reproduction.

### LWBC

In-Distribution Accuracy: 88.9. Result ID: 6990afee3a6c3148490275f0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Implementation](<https://github.com/nayeong-v-kim/LWBC>)

Recorded evaluation configuration: {"Backbone network":"Resnet18","Spurious attribute label":false}. This record does not assert independent reproduction.

### JTT

In-Distribution Accuracy: 88. Result ID: 6990afee3a6c3148490275ee.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ade1e209131cea2aeeb.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":false}. This record does not assert independent reproduction.

### LfF

In-Distribution Accuracy: 86. Result ID: 6990afee3a6c3148490275ec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68c11e209131cea1e11f.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":false}. This record does not assert independent reproduction.

### CVaR DRO

In-Distribution Accuracy: 82.4. Result ID: 6990afee3a6c3148490275eb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f331e209131cea44667.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa57a11e209131ce9a71e6.md>)

Recorded evaluation configuration: {"Backbone network":"Resnet50","Spurious attribute label":false}. This record does not assert independent reproduction.

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