# Object Detection on LVIS base categories v1.0 (minival5k train)

Source: [Object Detection on LVIS base categories v1.0 (minival5k train)](<https://www.sota2.com/research/sota/object-detection-on-lvis-base-categories-v1-0-minival5k-train>)

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

Data updated: 2026-02-26T05:23:02.822Z

Evaluation of Object Detection performance using the base categories subset of LVIS v1.0 on the minival5k training split, reporting metrics like AP\_all and AP\_rare.

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) | LVIS base categories v1.0 (minival5k train) |
| Metric | AP (All) |
| 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

- [AP (All)](<https://www.sota2.com/research/sota/object-detection-on-lvis-base-categories-v1-0-minival5k-train.md?metric=AP%20(All)>): Higher is better; unit: PERCENT.

- [AP (Rare)](<https://www.sota2.com/research/sota/object-detection-on-lvis-base-categories-v1-0-minival5k-train.md?metric=AP%20(Rare)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AP (All) | Result date | Report paper |
| --- | --- | --- | --- |
| DetCLIPv3 | 60.5 | Not recorded | 68fa72db1e209131cea58f4c |
| OWL-ST+FT | 56.2 | Not recorded | 68fa72db1e209131cea58f4c |
| DetCLIPv3 | 54.3 | Not recorded | 68fa72db1e209131cea58f4c |
| OWL-ST+FT | 48.7 | Not recorded | 68fa72db1e209131cea58f4c |

### DetCLIPv3

AP (All): 60.5. Result ID: 69929def3a6c314849102f2b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72db1e209131cea58f4c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72db1e209131cea58f4c.md>)

Recorded evaluation configuration: {"Backbone":"Swin-L","Fine-tuning protocol":"base categories only"}. This record does not assert independent reproduction.

### OWL-ST+FT

AP (All): 56.2. Result ID: 69929def3a6c314849102f2a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72db1e209131cea58f4c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b473.md>)

Recorded evaluation configuration: {"Backbone":"CLIP L/14","Fine-tuning protocol":"base categories only"}. This record does not assert independent reproduction.

### DetCLIPv3

AP (All): 54.3. Result ID: 69929def3a6c314849102f29.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72db1e209131cea58f4c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72db1e209131cea58f4c.md>)

Recorded evaluation configuration: {"Backbone":"Swin-T","Fine-tuning protocol":"base categories only"}. This record does not assert independent reproduction.

### OWL-ST+FT

AP (All): 48.7. Result ID: 69929def3a6c314849102f28.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72db1e209131cea58f4c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73401e209131cea5b473.md>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16","Fine-tuning protocol":"base categories only"}. This record does not assert independent reproduction.

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