# Image Classification on MNIST (Accuracy, NLL, ECE)

Source: [Image Classification on MNIST (Accuracy, NLL, ECE)](<https://www.sota2.com/research/sota/image-classification-on-mnist-accuracy-nll-ece>)

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

Data updated: 2026-02-26T05:26:47.266Z

Evaluation of standard Image Classification performance on the MNIST dataset, reporting Accuracy (ACC), Negative Log-Likelihood (NLL), and Expected Calibration Error (ECE).

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) | MNIST |
| Metric | Acc |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 13 of 13 |
| 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](<https://www.sota2.com/research/sota/image-classification-on-mnist-accuracy-nll-ece.md?metric=Acc>): Higher is better; unit: PERCENT.

- [NLL](<https://www.sota2.com/research/sota/image-classification-on-mnist-accuracy-nll-ece.md?metric=NLL>): Lower is better; unit: SCALAR.

- [ECE](<https://www.sota2.com/research/sota/image-classification-on-mnist-accuracy-nll-ece.md?metric=ECE>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Acc | Result date | Report paper |
| --- | --- | --- | --- |
| LLA | 97.9 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| Ensemble | 97.9 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| MAP | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| LLA | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| LLA | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| ELLA | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| VaLLA | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| VaLLA | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| GPP | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| Dropout | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| GAPA | 97.8 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| Linear Probing | 97.7 | Not recorded | 6993f6ab9ebc499c97409fe7 |
| LLA | 97.6 | Not recorded | 6993f6ab9ebc499c97409fe7 |

### LLA

Acc: 97.9. Result ID: 699a9dcf5698665ec4298cb7.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: {"Variant":"last-layer","Approximation":"KFAC"}. This record does not assert independent reproduction.

### Ensemble

Acc: 97.9. Result ID: 699a9dcf5698665ec4298cbe.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63bc1e209131cea0084f.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### MAP

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cb3.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65151e209131cea0817c.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLA

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cb5.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: {"Approximation":"KFAC"}. This record does not assert independent reproduction.

### LLA

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cb6.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: {"Variant":"last-layer"}. This record does not assert independent reproduction.

### ELLA

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cb8.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### VaLLA

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cb9.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: {"Inducing points":100}. This record does not assert independent reproduction.

### VaLLA

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cba.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: {"Inducing points":200}. This record does not assert independent reproduction.

### GPP

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cbc.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Dropout

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cbd.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62d31e209131ce9fb5ff.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GAPA

Acc: 97.8. Result ID: 699a9dcf5698665ec4298cbf.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Linear Probing

Acc: 97.7. Result ID: 699a9dcf5698665ec4298cbb.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65151e209131cea0817c.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLA

Acc: 97.6. Result ID: 699a9dcf5698665ec4298cb4.

[Reporting paper](<https://www.sota2.com/research/paper/6993f6ab9ebc499c97409fe7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b45b.md>)

Recorded evaluation configuration: {"Approximation":"Diagonal"}. This record does not assert independent reproduction.

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