Execution time modeling on Gated recurrent unit (GRU) layers on Galaxy Nexus (test)
2.9MAPEFastDeepIoT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FastDeepIoTHardware=Galaxy Nexus, Input Feature=explanatory variable vector x2018.09 | 2.9 | 1.1 | 0.997 | |
| DNNHardware=Galaxy Nexus, Input Feature=explanatory variable vector x, Layers=four-layer fully connected2018.09 | 4.1 | 2.2 | 0.989 | |
| GBRTHardware=Galaxy Nexus, Input Feature=explanatory variable vector x2018.09 | 6 | 3.2 | 0.984 | |
| RFHardware=Galaxy Nexus, Input Feature=explanatory variable vector x2018.09 | 8.8 | 4.1 | 0.977 | |
| DTHardware=Galaxy Nexus, Input Feature=explanatory variable vector x2018.09 | 10.5 | 4.8 | 0.968 | |
| SVRHardware=Galaxy Nexus, Input Feature=explanatory variable vector x, Kernel=radial basis function2018.09 | 71.5 | 27.8 | -0.065 |