# Object Detection on ADE20K open-vocabulary

Source: [Object Detection on ADE20K open-vocabulary](<https://www.sota2.com/research/sota/object-detection-on-ade20k-open-vocabulary>)

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

Data updated: 2026-02-26T05:23:03.572Z

Evaluation of Object Detection performance on the open-vocabulary setting of the ADE20K dataset, typically measured using bounding box Average Precision (box AP).

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 | Object Detection |
| Dataset (including any stated split) | ADE20K open-vocabulary |
| Metric | AP (Box) |
| 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

- [AP (Box)](<https://www.sota2.com/research/sota/object-detection-on-ade20k-open-vocabulary.md?metric=AP%20(Box)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AP (Box) | Result date | Report paper |
| --- | --- | --- | --- |
| OpenSD | 17.8 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSeeD | 17 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSD | 16.8 | Not recorded | 68fa717b1e209131cea51877 |
| HIPIE | 16.2 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSD | 15.3 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSD | 12.2 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSeeD | 11.5 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSD | 11.4 | Not recorded | 68fa717b1e209131cea51877 |
| OpenSeeD | 10.7 | Not recorded | 68fa717b1e209131cea51877 |

### OpenSD

AP (Box): 17.8. Result ID: 6991d2293a6c3148490acd38.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Implementation](<https://github.com/strongwolf/OpenSD>)

Recorded evaluation configuration: {"Training Data":"COCO+O365","Backbone":"Swin-T"}. This record does not assert independent reproduction.

### OpenSeeD

AP (Box): 17. Result ID: 6991d2293a6c3148490acd37.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ecf1e209131cea42550.md>)

[Implementation](<https://github.com/IDEA-Research/OpenSeeD>)

Recorded evaluation configuration: {"Training Data":"COCO+O365","Backbone":"Swin-T"}. This record does not assert independent reproduction.

### OpenSD

AP (Box): 16.8. Result ID: 6991d2293a6c3148490acd34.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Implementation](<https://github.com/strongwolf/OpenSD>)

Recorded evaluation configuration: {"Training Data":"COCO+O365","Backbone":"R50"}. This record does not assert independent reproduction.

### HIPIE

AP (Box): 16.2. Result ID: 6991d2293a6c3148490acd32.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71a81e209131cea525cd.md>)

[Implementation](<http://people.eecs.berkeley.edu/~similar-to/simxdwang/projects/HIPIE>)

Recorded evaluation configuration: {"Training Data":"COCO+O365","Backbone":"R50"}. This record does not assert independent reproduction.

### OpenSD

AP (Box): 15.3. Result ID: 6991d2293a6c3148490acd3a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Implementation](<https://github.com/strongwolf/OpenSD>)

Recorded evaluation configuration: {"Training Data":"COCO","Backbone":"Swin-L"}. This record does not assert independent reproduction.

### OpenSD

AP (Box): 12.2. Result ID: 6991d2293a6c3148490acd36.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Implementation](<https://github.com/strongwolf/OpenSD>)

Recorded evaluation configuration: {"Training Data":"COCO","Backbone":"Swin-T"}. This record does not assert independent reproduction.

### OpenSeeD

AP (Box): 11.5. Result ID: 6991d2293a6c3148490acd39.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ecf1e209131cea42550.md>)

[Implementation](<https://github.com/IDEA-Research/OpenSeeD>)

Recorded evaluation configuration: {"Training Data":"COCO","Backbone":"Swin-L"}. This record does not assert independent reproduction.

### OpenSD

AP (Box): 11.4. Result ID: 6991d2293a6c3148490acd33.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Implementation](<https://github.com/strongwolf/OpenSD>)

Recorded evaluation configuration: {"Training Data":"COCO","Backbone":"R50"}. This record does not assert independent reproduction.

### OpenSeeD

AP (Box): 10.7. Result ID: 6991d2293a6c3148490acd35.

[Reporting paper](<https://www.sota2.com/research/paper/68fa717b1e209131cea51877.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ecf1e209131cea42550.md>)

[Implementation](<https://github.com/IDEA-Research/OpenSeeD>)

Recorded evaluation configuration: {"Training Data":"COCO","Backbone":"Swin-T"}. This record does not assert independent reproduction.

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