# Image Classification on SUN397 (16-shot)

Source: [Image Classification on SUN397 (16-shot)](<https://www.sota2.com/research/sota/image-classification-on-sun397-16-shot>)

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

Data updated: 2026-04-03T07:34:10.847Z

Evaluation of Image Classification performance on the SUN397 dataset using a 16-shot few-shot learning protocol, typically measured by accuracy.

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) | SUN397 16-shot |
| Metric | Accuracy |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/image-classification-on-sun397-16-shot.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| VR (glass-box) | 66.7 | Not recorded | 69cf49700636e5bf4afcd91e |
| BlackVIP | 64.5 | Not recorded | 69cf49700636e5bf4afcd91e |
| AReS | 62.8 | Not recorded | 69cf49700636e5bf4afcd91e |
| AReS-MS | 62.8 | Not recorded | 69cf49700636e5bf4afcd91e |
| Zero-shot | 62.6 | Not recorded | 69cf49700636e5bf4afcd91e |
| BAR | 62.4 | Not recorded | 69cf49700636e5bf4afcd91e |
| LLM-Opt | 60.3 | Not recorded | 69cf49700636e5bf4afcd91e |

### VR (glass-box)

Accuracy: 66.7. Result ID: 69cf6c4d47f34ce924b8ea10.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75741e209131cea68084.md>)

[Implementation](<https://github.com/tmlr-group/BayesianLM>)

Recorded evaluation configuration: {"Backbone":"CLIP ViT-B/16 (Service)","Shot Setting":"16-shot","#API (M)":"16.82","Time (h)":"9.8"}. This record does not assert independent reproduction.

### BlackVIP

Accuracy: 64.5. Result ID: 69cf6c4d47f34ce924b8ea13.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ffb1e209131cea484ba.md>)

[Implementation](<https://github.com/changdaeoh/BlackVIP>)

Recorded evaluation configuration: {"Backbone":"CLIP ViT-B/16 (Service)","Shot Setting":"16-shot","#API (M)":"754.20","Time (h)":"197.5"}. This record does not assert independent reproduction.

### AReS

Accuracy: 62.8. Result ID: 69cf6c4d47f34ce924b8ea15.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (Local)","Shot Setting":"16-shot","#API (M)":"0.02","Time (h)":"3.7"}. This record does not assert independent reproduction.

### AReS-MS

Accuracy: 62.8. Result ID: 69cf6c4d47f34ce924b8ea16.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

Recorded evaluation configuration: {"Backbone":"Service & Local selection","Shot Setting":"16-shot","#API (M)":"0.06","Time (h)":"3.7"}. This record does not assert independent reproduction.

### Zero-shot

Accuracy: 62.6. Result ID: 69cf6c4d47f34ce924b8ea11.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

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

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

Recorded evaluation configuration: {"Backbone":"CLIP ViT-B/16 (Service)","Shot Setting":"16-shot","#API (M)":"0.12","Time (h)":"0"}. This record does not assert independent reproduction.

### BAR

Accuracy: 62.4. Result ID: 69cf6c4d47f34ce924b8ea12.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68081e209131cea19ab5.md>)

Recorded evaluation configuration: {"Backbone":"CLIP ViT-B/16 (Service)","Shot Setting":"16-shot","#API (M)":"612.84","Time (h)":"185.6"}. This record does not assert independent reproduction.

### LLM-Opt

Accuracy: 60.3. Result ID: 69cf6c4d47f34ce924b8ea14.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49700636e5bf4afcd91e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73331e209131cea5b19f.md>)

[Implementation](<https://www.sota2.com/llm-can-optimize-vlm.github.io>)

Recorded evaluation configuration: {"Backbone":"CLIP ViT-B/16 (Service)","Shot Setting":"16-shot","#API (M)":"5011.32","Time (h)":"5.1"}. This record does not assert independent reproduction.

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