Nash-MTL

Reported on 6 benchmarks across 2 tasks · 1 paper · 6 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology6 results

Transfer LearningonNYUv2
Mean IoU· 2022-02-02
40.13
best: 58.14 (SwinMTL)
SOTA
Multi-Task Learning as a Bargaining Game arXiv:2202.01017
Transfer LearningonCityscapes test
mIoU· 2022-02-02
75.41
best: 76.41 (SwinMTL)
SOTA
Multi-Task Learning as a Bargaining Game arXiv:2202.01017
Transfer LearningonQM9
∆m%· 2022-02-02
62
best: 125.7 (PCGrad)
SOTA
Multi-Task Learning as a Bargaining Game arXiv:2202.01017
Multi-Task LearningonNYUv2
Mean IoU· 2022-02-02
40.13
best: 58.14 (SwinMTL)
SOTA
Multi-Task Learning as a Bargaining Game arXiv:2202.01017
Multi-Task LearningonCityscapes test
mIoU· 2022-02-02
75.41
best: 76.41 (SwinMTL)
SOTA
Multi-Task Learning as a Bargaining Game arXiv:2202.01017
Multi-Task LearningonQM9
∆m%· 2022-02-02
62
best: 125.7 (PCGrad)
SOTA
Multi-Task Learning as a Bargaining Game arXiv:2202.01017