# Panoptic Segmentation on ADE20K 150 (val)

Source: [Panoptic Segmentation on ADE20K 150 (val)](<https://www.sota2.com/research/sota/panoptic-segmentation-on-ade20k-150-val>)

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

Data updated: 2026-05-19T09:03:04.578Z

Evaluation of Panoptic Segmentation performance on the validation split of the ADE20K dataset, specifically using the 150-class configuration.

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 | Panoptic Segmentation |
| Dataset (including any stated split) | ADE20K 150 (val) |
| Metric | PQ |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 18 of 18 |
| 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

- [PQ](<https://www.sota2.com/research/sota/panoptic-segmentation-on-ade20k-150-val.md?metric=PQ>): Higher is better; unit: PERCENT.

- [mAP](<https://www.sota2.com/research/sota/panoptic-segmentation-on-ade20k-150-val.md?metric=mAP>): Higher is better; unit: PERCENT.

- [mIoU](<https://www.sota2.com/research/sota/panoptic-segmentation-on-ade20k-150-val.md?metric=mIoU>): Higher is better; unit: PERCENT.

## Recorded results

| Method | PQ | Result date | Report paper |
| --- | --- | --- | --- |
| WOW-Seg | 44.46 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| Osprey-7B | 41.89 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| Ferret-7B | 39.46 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| CLIP-ConvNeXt-L | 36.86 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| GPT4RoI-7B | 36.32 | Not recorded | 68fa794e1e209131cea7d037 |
| GPT4RoI-7B | 36.32 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| RetCLIP | 30.9 | Not recorded | 69705d93e25ad2fa9d1ed152 |
| RetCLIP | 28.3 | Not recorded | 69705d93e25ad2fa9d1ed152 |
| Shikra-7B | 27.52 | Not recorded | 68fa794e1e209131cea7d037 |
| Shikra-7B | 27.52 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| RetCLIP | 27.1 | Not recorded | 69705d93e25ad2fa9d1ed152 |
| CLIP-Surgery-ViT-L | 26.55 | Not recorded | 68fa794e1e209131cea7d037 |
| CLIP-Surgery-ViT-L | 26.55 | Not recorded | 6a0beeb6dbc836b87d82c559 |
| FC-CLIP | 26.4 | Not recorded | 69705d93e25ad2fa9d1ed152 |
| RetCLIP | 23.7 | Not recorded | 69705d93e25ad2fa9d1ed152 |
| FC-CLIP | 22.8 | Not recorded | 69705d93e25ad2fa9d1ed152 |
| Kosmos-2 | 6.53 | Not recorded | 68fa794e1e209131cea7d037 |
| Kosmos-2 | 6.53 | Not recorded | 6a0beeb6dbc836b87d82c559 |

### WOW-Seg

PQ: 44.46. Result ID: 6a0c273cfbc373592c026694.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### Osprey-7B

PQ: 41.89. Result ID: 6a0c273cfbc373592c026693.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b1f1e209131cea86abf.md>)

[Implementation](<https://github.com/CircleRadon/Osprey>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### Ferret-7B

PQ: 39.46. Result ID: 6a0c273cfbc373592c026692.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d81e209131cea4da64.md>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### CLIP-ConvNeXt-L

PQ: 36.86. Result ID: 6a0c273cfbc373592c02668d.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

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

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

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### GPT4RoI-7B

PQ: 36.32. Result ID: 6995d9003badfaffe6df3294.

[Reporting paper](<https://www.sota2.com/research/paper/68fa794e1e209131cea7d037.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa794e1e209131cea7d037.md>)

[Implementation](<https://github.com/jshilong/GPT4RoI>)

Recorded evaluation configuration: {"Model size":"7B"}. This record does not assert independent reproduction.

### GPT4RoI-7B

PQ: 36.32. Result ID: 6a0c273cfbc373592c026691.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa794e1e209131cea7d037.md>)

[Implementation](<https://github.com/jshilong/GPT4RoI>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### RetCLIP

PQ: 30.9. Result ID: 6997ab8a5698665ec401af96.

[Reporting paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-ConvNeXt-large","Database":"ADE20k","Fine-tuned on COCO":true}. This record does not assert independent reproduction.

### RetCLIP

PQ: 28.3. Result ID: 6997ab8a5698665ec401af95.

[Reporting paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-ConvNeXt-large","Database":"Google Open Image","Fine-tuned on COCO":true}. This record does not assert independent reproduction.

### Shikra-7B

PQ: 27.52. Result ID: 6995d9003badfaffe6df3293.

[Reporting paper](<https://www.sota2.com/research/paper/68fa794e1e209131cea7d037.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fe61e209131cea47fbc.md>)

[Implementation](<https://github.com/shikras/shikra>)

Recorded evaluation configuration: {"Concurrent work":"true","Model size":"7B"}. This record does not assert independent reproduction.

### Shikra-7B

PQ: 27.52. Result ID: 6a0c273cfbc373592c026690.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fe61e209131cea47fbc.md>)

[Implementation](<https://github.com/shikras/shikra>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### RetCLIP

PQ: 27.1. Result ID: 6997ab8a5698665ec401af93.

[Reporting paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50x64","Database":"ADE20k","Fine-tuned on COCO":true}. This record does not assert independent reproduction.

### CLIP-Surgery-ViT-L

PQ: 26.55. Result ID: 6995d9003badfaffe6df3291.

[Reporting paper](<https://www.sota2.com/research/paper/68fa794e1e209131cea7d037.md>)

Recorded evaluation configuration: {"Backbone":"ViT-L"}. This record does not assert independent reproduction.

### CLIP-Surgery-ViT-L

PQ: 26.55. Result ID: 6a0c273cfbc373592c02668e.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

### FC-CLIP

PQ: 26.4. Result ID: 6997ab8a5698665ec401af94.

[Reporting paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71291e209131cea4f931.md>)

[Implementation](<https://github.com/bytedance/fc-clip>)

Recorded evaluation configuration: {"Backbone":"CLIP-ConvNeXt-large","Database":"N/A","Fine-tuned on COCO":true}. This record does not assert independent reproduction.

### RetCLIP

PQ: 23.7. Result ID: 6997ab8a5698665ec401af92.

[Reporting paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50x64","Database":"Google Open Image","Fine-tuned on COCO":true}. This record does not assert independent reproduction.

### FC-CLIP

PQ: 22.8. Result ID: 6997ab8a5698665ec401af91.

[Reporting paper](<https://www.sota2.com/research/paper/69705d93e25ad2fa9d1ed152.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71291e209131cea4f931.md>)

[Implementation](<https://github.com/bytedance/fc-clip>)

Recorded evaluation configuration: {"Backbone":"CLIP-RN50x64","Database":"N/A","Fine-tuned on COCO":true}. This record does not assert independent reproduction.

### Kosmos-2

PQ: 6.53. Result ID: 6995d9003badfaffe6df3292.

[Reporting paper](<https://www.sota2.com/research/paper/68fa794e1e209131cea7d037.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa700a1e209131cea48c7f.md>)

[Implementation](<https://aka.ms/kosmos-2>)

Recorded evaluation configuration: {"Concurrent work":"true"}. This record does not assert independent reproduction.

### Kosmos-2

PQ: 6.53. Result ID: 6a0c273cfbc373592c02668f.

[Reporting paper](<https://www.sota2.com/research/paper/6a0beeb6dbc836b87d82c559.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa700a1e209131cea48c7f.md>)

[Implementation](<https://aka.ms/kosmos-2>)

Recorded evaluation configuration: {"Protocol":"Open-vocabulary","Evaluation Input":"Ground truth box/mask"}. This record does not assert independent reproduction.

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