# Semantic Segmentation on ADE20K (val) (Partitioned mIoU)

Source: [Semantic Segmentation on ADE20K (val) (Partitioned mIoU)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val-partitioned-miou>)

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

Data updated: 2026-02-26T06:10:53.398Z

Evaluation of Semantic Segmentation performance on the validation split of the ADE20K dataset, specifically reporting Mean IoU metrics calculated across various input scale partitions (1/32, 1/16, 1/8, 1/4, 1/2).

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 | Semantic Segmentation |
| Dataset (including any stated split) | ADE20K (val) |
| Metric | mIoU (1/32 partition) |
| 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

- [mIoU (1/32 partition)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val-partitioned-miou.md?metric=mIoU%20(1%2F32%20partition)>): Higher is better; unit: PERCENT.

- [mIoU (1/16 partition)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val-partitioned-miou.md?metric=mIoU%20(1%2F16%20partition)>): Higher is better; unit: PERCENT.

- [mIoU (1/8 partition)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val-partitioned-miou.md?metric=mIoU%20(1%2F8%20partition)>): Higher is better; unit: PERCENT.

- [mIoU (1/4 partition)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val-partitioned-miou.md?metric=mIoU%20(1%2F4%20partition)>): Higher is better; unit: PERCENT.

- [mIoU (1/2 partition)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val-partitioned-miou.md?metric=mIoU%20(1%2F2%20partition)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | mIoU (1/32 partition) | Result date | Report paper |
| --- | --- | --- | --- |
| Dual Teacher | 22.6 | Not recorded | 68fa71041e209131cea4e95a |
| ReCo | 21.11 | Not recorded | 68fa71041e209131cea4e95a |
| GTA-Seg | 19.52 | Not recorded | 68fa71041e209131cea4e95a |
| CPS | 18.79 | Not recorded | 68fa71041e209131cea4e95a |
| U2PL | 18.54 | Not recorded | 68fa71041e209131cea4e95a |
| Supervised-only | 15.15 | Not recorded | 68fa71041e209131cea4e95a |

### Dual Teacher

mIoU (1/32 partition): 22.6. Result ID: 699189993a6c31484908ce72.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

[Implementation](<https://github.com/naver-ai/dual-teacher>)

Recorded evaluation configuration: {"#Params":"13.7M","Backbone":"SegFormer with MiT-B1","Inference Scale":"single-scale"}. This record does not assert independent reproduction.

### ReCo

mIoU (1/32 partition): 21.11. Result ID: 699189993a6c31484908ce71.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bba1e209131cea2fee1.md>)

[Implementation](<https://github.com/lorenmt/reco>)

Recorded evaluation configuration: {"#Params":"14.4M","Backbone":"SegFormer with MiT-B1","Inference Scale":"single-scale","re-implemented":true}. This record does not assert independent reproduction.

### GTA-Seg

mIoU (1/32 partition): 19.52. Result ID: 699189993a6c31484908ce70.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e641e209131cea3fa7c.md>)

[Implementation](<https://github.com/Jin-Ying/GTA-Seg>)

Recorded evaluation configuration: {"#Params":"28.8M","Backbone":"SegFormer with MiT-B1","Inference Scale":"single-scale","re-implemented":true}. This record does not assert independent reproduction.

### CPS

mIoU (1/32 partition): 18.79. Result ID: 699189993a6c31484908ce6e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a231e209131cea26482.md>)

[Implementation](<https://git.io/CPS>)

Recorded evaluation configuration: {"#Params":"27.5M","Backbone":"SegFormer with MiT-B1","Inference Scale":"single-scale","re-implemented":true}. This record does not assert independent reproduction.

### U2PL

mIoU (1/32 partition): 18.54. Result ID: 699189993a6c31484908ce6f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bff1e209131cea318a4.md>)

[Implementation](<https://haochen-wang409.github.io/U2PL>)

Recorded evaluation configuration: {"#Params":"15.0M","Backbone":"SegFormer with MiT-B1","Inference Scale":"single-scale","re-implemented":true}. This record does not assert independent reproduction.

### Supervised-only

mIoU (1/32 partition): 15.15. Result ID: 699189993a6c31484908ce6d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71041e209131cea4e95a.md>)

Recorded evaluation configuration: {"#Params":"13.7M","Backbone":"SegFormer with MiT-B1","Inference Scale":"single-scale"}. This record does not assert independent reproduction.

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