Multi-Task Learning 벤치마크
Multi-Task Learning on QM9
∆m%
- 2022-02-02 — Nash-MTL: ∆m% 62.0
| Rank | Model | ∆m% | Paper | Code | Year |
|---|---|---|---|---|---|
| 1 | BayesAgg-MTL | 53.7 | Bayesian Uncertainty for Gradient Aggregation in Multi-Task Learning | ssi-research/bayesagg_mtl | 2024 |
| 2 | Nash-MTL | 62.0 | Multi-Task Learning as a Bargaining Game | torchjd/torchjd · avivnavon/nash-mtl · cranial-xix/famo · +1 | 2022 |
| 3 | IMTL-G | 77.2 | |||
| 4 | CAGrad | 112.8 | |||
| 5 | PCGrad | 125.7 | |||
| 6 | BayesAgg-MTL | 53.7 | Bayesian Uncertainty for Gradient Aggregation in Multi-Task Learning | ssi-research/bayesagg_mtl | 2024 |
| 7 | Nash-MTL | 62.0 | Multi-Task Learning as a Bargaining Game | torchjd/torchjd · avivnavon/nash-mtl · cranial-xix/famo · +1 | 2022 |
| 8 | IMTL-G | 77.2 | |||
| 9 | CAGrad | 112.8 | |||
| 10 | PCGrad | 125.7 |