Torque 7 Prediction on SARCOS robot arm (test)
0.79RMSEFine-tune
Evaluation results
| Method | Links | |
|---|---|---|
| Fine-tuneNumber of training samples=502022.10 | 0.79 | |
| L2-SPNumber of training samples=502022.10 | 0.802 | |
| Fine-tuneNumber of training samples=402022.10 | 0.848 | |
| L2-SPNumber of training samples=402022.10 | 0.858 | |
| AffineTL-fullNumber of training samples=502022.10 | 0.87 | |
| AffineTL-constNumber of training samples=502022.10 | 0.884 | |
| AugmentedNumber of training samples=502022.10 | 0.897 | |
| HTL-offsetNumber of training samples=502022.10 | 0.925 | |
| Only sourceNumber of training samples=502022.10 | 0.927 | |
| HTL-scaleNumber of training samples=502022.10 | 0.948 | |
| Fine-tuneNumber of training samples=302022.10 | 0.959 | |
| AffineTL-constNumber of training samples=402022.10 | 0.963 | |
| L2-SPNumber of training samples=302022.10 | 0.968 | |
| Only sourceNumber of training samples=402022.10 | 0.969 | |
| HTL-offsetNumber of training samples=402022.10 | 0.969 | |
| AffineTL-fullNumber of training samples=402022.10 | 0.974 | |
| AugmentedNumber of training samples=402022.10 | 0.984 | |
| HTL-scaleNumber of training samples=402022.10 | 0.999 | |
| AffineTL-constNumber of training samples=302022.10 | 1.04 | |
| DirectNumber of training samples=502022.10 | 1.05 | |
| AffineTL-fullNumber of training samples=302022.10 | 1.06 | |
| MAMLNumber of training samples=502022.10 | 1.06 | |
| Only sourceNumber of training samples=302022.10 | 1.09 | |
| PAC-NetNumber of training samples=502022.10 | 1.1 | |
| HTL-scaleNumber of training samples=302022.10 | 1.12 | |
| Fine-tuneNumber of training samples=202022.10 | 1.12 | |
| L2-SPNumber of training samples=202022.10 | 1.12 | |
| PAC-NetNumber of training samples=402022.10 | 1.13 | |
| AugmentedNumber of training samples=302022.10 | 1.16 | |
| DirectNumber of training samples=402022.10 | 1.18 | |
| PAC-NetNumber of training samples=302022.10 | 1.18 | |
| AffineTL-constNumber of training samples=202022.10 | 1.2 | |
| MAMLNumber of training samples=402022.10 | 1.2 | |
| AffineTL-fullNumber of training samples=202022.10 | 1.21 | |
| Only sourceNumber of training samples=202022.10 | 1.22 | |
| HTL-offsetNumber of training samples=302022.10 | 1.22 | |
| HTL-scaleNumber of training samples=202022.10 | 1.24 | |
| PAC-NetNumber of training samples=202022.10 | 1.24 | |
| MAMLNumber of training samples=302022.10 | 1.28 | |
| AugmentedNumber of training samples=202022.10 | 1.31 | |
| Fine-tuneNumber of training samples=152022.10 | 1.35 | |
| L2-SPNumber of training samples=152022.10 | 1.35 | |
| PAC-NetNumber of training samples=152022.10 | 1.43 | |
| AffineTL-fullNumber of training samples=152022.10 | 1.45 | |
| AffineTL-constNumber of training samples=152022.10 | 1.48 | |
| Only sourceNumber of training samples=152022.10 | 1.49 | |
| HTL-offsetNumber of training samples=202022.10 | 1.49 | |
| HTL-scaleNumber of training samples=152022.10 | 1.51 | |
| DirectNumber of training samples=302022.10 | 1.54 | |
| PAC-NetNumber of training samples=102022.10 | 1.61 | |
| Fine-tuneNumber of training samples=102022.10 | 1.62 | |
| MAMLNumber of training samples=202022.10 | 1.63 | |
| L2-SPNumber of training samples=102022.10 | 1.65 | |
| AugmentedNumber of training samples=152022.10 | 1.67 | |
| MAMLNumber of training samples=152022.10 | 1.67 | |
| HTL-offsetNumber of training samples=152022.10 | 1.69 | |
| HTL-scaleNumber of training samples=102022.10 | 1.71 | |
| AffineTL-constNumber of training samples=102022.10 | 1.71 | |
| Only sourceNumber of training samples=102022.10 | 1.73 | |
| AffineTL-fullNumber of training samples=102022.10 | 1.73 | |
| DirectNumber of training samples=202022.10 | 1.85 | |
| HTL-offsetNumber of training samples=102022.10 | 1.85 | |
| DirectNumber of training samples=152022.10 | 1.9 | |
| AugmentedNumber of training samples=102022.10 | 1.9 | |
| MAMLNumber of training samples=102022.10 | 1.9 | |
| Only sourceNumber of training samples=52022.10 | 2.11 | |
| DirectNumber of training samples=102022.10 | 2.13 | |
| AffineTL-fullNumber of training samples=52022.10 | 2.23 | |
| PAC-NetNumber of training samples=52022.10 | 2.24 | |
| HTL-offsetNumber of training samples=52022.10 | 2.29 | |
| AffineTL-constNumber of training samples=52022.10 | 2.3 | |
| HTL-scaleNumber of training samples=52022.10 | 2.32 | |
| Fine-tuneNumber of training samples=52022.10 | 2.33 | |
| L2-SPNumber of training samples=52022.10 | 2.33 | |
| AugmentedNumber of training samples=52022.10 | 2.41 | |
| MAMLNumber of training samples=52022.10 | 2.54 | |
| DirectNumber of training samples=52022.10 | 2.66 |