Visual Domain Classification on Visual Domain Decathlon Benchmark (VDDB) (test)
85.56Max Test Accuracyμ2Net+
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| μ2Net+2022.09 | 85.56 | 85.19 | 0.37 | 68.7 | 7,001 | |
| muNet2022.09 | 83.58 | 83.2 | 0.29 | 45 | 5,490 | |
| Fine-tuningfrozen_layers=below 1st transformer layer2022.09 | 81.51 | 81.39 | 0.14 | 78.98 | 4,485 | |
| Residual adaptersdimension=2562022.09 | 80.67 | 80.58 | 0.07 | 13.56 | 4,734 |