# Image Classification on FMNIST (test) (Accuracy, NLL, ECE)

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

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

Data updated: 2026-05-18T07:17:47.526Z

Evaluation of Image Classification performance on the test split of the Fashion MNIST dataset, reporting standard accuracy alongside calibration metrics including Negative Log-Likelihood (NLL), Brier Score, Expected Calibration Error (ECE), and Maximum Calibration Error (MCE).

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) | FMNIST (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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

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

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

- [Brier Score](<https://www.sota2.com/research/sota/image-classification-on-fmnist-test-accuracy-nll-ece.md?metric=Brier%20Score>): Lower is better; unit: SCALAR.

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

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

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| DIMS | 81.4 | Not recorded | 6a0a9d15dbc836b87d82b40d |
| RLA | 76.7 | Not recorded | 6a0a9d15dbc836b87d82b40d |
| LINLA | 38.5 | Not recorded | 6a0a9d15dbc836b87d82b40d |
| LA | 14.7 | Not recorded | 6a0a9d15dbc836b87d82b40d |

### DIMS

Accuracy: 81.4. Result ID: 6a0abbfc47f34ce924f067aa.

[Reporting paper](<https://www.sota2.com/research/paper/6a0a9d15dbc836b87d82b40d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0a9d15dbc836b87d82b40d.md>)

Recorded evaluation configuration: {"Hessian approximation type":"low-rank","Backbone":"LeNet","Number of parameter samples":"100","friction coefficient η0":"0.1"}. This record does not assert independent reproduction.

### RLA

Accuracy: 76.7. Result ID: 6a0abbfc47f34ce924f067a9.

[Reporting paper](<https://www.sota2.com/research/paper/6a0a9d15dbc836b87d82b40d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6faf1e209131cea46e23.md>)

Recorded evaluation configuration: {"Hessian approximation type":"low-rank","Backbone":"LeNet","Number of parameter samples":"100"}. This record does not assert independent reproduction.

### LINLA

Accuracy: 38.5. Result ID: 6a0abbfc47f34ce924f067a8.

[Reporting paper](<https://www.sota2.com/research/paper/6a0a9d15dbc836b87d82b40d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69541e209131cea2192f.md>)

Recorded evaluation configuration: {"Hessian approximation type":"low-rank","Backbone":"LeNet","Number of parameter samples":"100"}. This record does not assert independent reproduction.

### LA

Accuracy: 14.7. Result ID: 6a0abbfc47f34ce924f067a7.

[Reporting paper](<https://www.sota2.com/research/paper/6a0a9d15dbc836b87d82b40d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6faf1e209131cea46e23.md>)

Recorded evaluation configuration: {"Hessian approximation type":"low-rank","Backbone":"LeNet","Number of parameter samples":"100"}. This record does not assert independent reproduction.

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