# Image Classification on RESISC45 (Accuracy and Calibration)

Source: [Image Classification on RESISC45 (Accuracy and Calibration)](<https://www.sota2.com/research/sota/image-classification-on-resisc45-accuracy-and-calibration>)

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

Data updated: 2026-06-23T12:32:00.561Z

Evaluation of Image Classification performance on the RESISC45 dataset, focusing on standard accuracy alongside model calibration metrics like NLL, ECE, and Brier Score.

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) | RESISC45 |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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-resisc45-accuracy-and-calibration.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [NLL](<https://www.sota2.com/research/sota/image-classification-on-resisc45-accuracy-and-calibration.md?metric=NLL>): Lower is better; unit: SCALAR.

- [ECE](<https://www.sota2.com/research/sota/image-classification-on-resisc45-accuracy-and-calibration.md?metric=ECE>): Lower is better; unit: PERCENT.

- [Brier Score](<https://www.sota2.com/research/sota/image-classification-on-resisc45-accuracy-and-calibration.md?metric=Brier%20Score>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| EVON | 95.79 | Not recorded | 6a3a1392206542251d7c7697 |
| SOAP | 95.74 | Not recorded | 6a3a1392206542251d7c7697 |
| AdamW | 95.61 | Not recorded | 6a3a1392206542251d7c7697 |
| EVON | 95.31 | Not recorded | 6a3a1392206542251d7c7697 |
| IVON | 95.23 | Not recorded | 6a3a1392206542251d7c7697 |
| IVON | 95.17 | Not recorded | 6a3a1392206542251d7c7697 |

### EVON

Accuracy: 95.79. Result ID: 6a3a7ca6fbc373592c38ed4a.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

Recorded evaluation configuration: {"Optimizer":"EVON","Monte Carlo samples (BMA)":"32","Backbone":"ViT-B/16","Pre-training":"OpenAI CLIP","Fine-tuning protocol":"linear layers (99% of parameters)"}. This record does not assert independent reproduction.

### SOAP

Accuracy: 95.74. Result ID: 6a3a7ca6fbc373592c38ed48.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

Recorded evaluation configuration: {"Optimizer":"SOAP","Monte Carlo samples (BMA)":"1","Backbone":"ViT-B/16","Pre-training":"OpenAI CLIP","Fine-tuning protocol":"linear layers (99% of parameters)"}. This record does not assert independent reproduction.

### AdamW

Accuracy: 95.61. Result ID: 6a3a7ca6fbc373592c38ed45.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

Recorded evaluation configuration: {"Optimizer":"AdamW","Monte Carlo samples (BMA)":"1","Backbone":"ViT-B/16","Pre-training":"OpenAI CLIP","Fine-tuning protocol":"linear layers (99% of parameters)"}. This record does not assert independent reproduction.

### EVON

Accuracy: 95.31. Result ID: 6a3a7ca6fbc373592c38ed49.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

Recorded evaluation configuration: {"Optimizer":"EVON","Monte Carlo samples (BMA)":"1","Backbone":"ViT-B/16","Pre-training":"OpenAI CLIP","Fine-tuning protocol":"linear layers (99% of parameters)"}. This record does not assert independent reproduction.

### IVON

Accuracy: 95.23. Result ID: 6a3a7ca6fbc373592c38ed47.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa5be81e209131ce9cc869.md>)

Recorded evaluation configuration: {"Optimizer":"IVON","Monte Carlo samples (BMA)":"32","Backbone":"ViT-B/16","Pre-training":"OpenAI CLIP","Fine-tuning protocol":"linear layers (99% of parameters)"}. This record does not assert independent reproduction.

### IVON

Accuracy: 95.17. Result ID: 6a3a7ca6fbc373592c38ed46.

[Reporting paper](<https://www.sota2.com/research/paper/6a3a1392206542251d7c7697.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa5be81e209131ce9cc869.md>)

Recorded evaluation configuration: {"Optimizer":"IVON","Monte Carlo samples (BMA)":"1","Backbone":"ViT-B/16","Pre-training":"OpenAI CLIP","Fine-tuning protocol":"linear layers (99% of parameters)"}. This record does not assert independent reproduction.

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