Image Classification on M(CIFAR100-20, F-CelebA)
74.68AccuracyCAT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| CATArchitecture=AlexNet-like, Number of random sequences=52021.12 | 74.68 | 73.39 | 77.27 | |
| ONEArchitecture=AlexNet-like, Number of random sequences=52021.12 | 74.28 | 78.7 | 65.45 | |
| ONENetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 70.58 | — | — | |
| HATArchitecture=AlexNet-like, Number of random sequences=52021.12 | 69.46 | 74.19 | 60 | |
| NCLArchitecture=AlexNet-like, Number of random sequences=52021.12 | 69.31 | 66.69 | 74.55 | |
| PRONetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 68.96 | — | — | |
| HATNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 68.02 | — | — | |
| UCLNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 66.89 | — | — | |
| CATNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 66.83 | — | — | |
| EWCNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 63.48 | — | — | |
| PathNetNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 61.63 | — | — | |
| NCLNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 61.36 | — | — | |
| HYPNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 60.53 | — | — | |
| RPSNetNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 55.07 | — | — | |
| HYP-RNetwork Architecture=2-layer fully connected network, Random Task Sequences=52021.12 | 32.74 | — | — |