Multi-task Learning on NYUD v2 (Am)
0AmIndependent
Evaluation results
| Method | Links | |
|---|---|---|
| IndependentBackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 0 | |
| MGDABackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 9.21 | |
| PCGradBackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 9.4 | |
| GDBackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 9.53 | |
| CAGradBackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 9.84 | |
| Aligned-MTLBackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 10.17 | |
| OursBackbone=HRNet-18, Task Weights=Manually tuned, Seeds=32024.06 | 15 |