Execution time modeling on Nexus 5 profiling Convolutional layer (test)
6.4MAPEDNN
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DNNHardware=Nexus 5, Layer Type=Convolutional2018.09 | 6.4 | 16.4 | 0.994 | |
| DNNArchitecture=four-layer fully connected neural network, Input features=explanatory variable vector x, Device=Nexus 5, Target layer=Convolutional layer2018.09 | 6.4 | 16.4 | 0.994 | |
| FastDeepIoTHardware=Nexus 5, Layer Type=Convolutional2018.09 | 7.6 | 15.2 | 0.991 | |
| FastDeepIoTModel type=tree-structured linear regression model, Input features=explanatory variable vector x, Device=Nexus 5, Target layer=Convolutional layer2018.09 | 7.6 | 15.2 | 0.991 | |
| GBRTHardware=Nexus 5, Layer Type=Convolutional2018.09 | 10.9 | 20.5 | 0.988 | |
| GBRTDescription=gradient boosted regression trees, Input features=explanatory variable vector x, Device=Nexus 5, Target layer=Convolutional layer2018.09 | 10.9 | 20.5 | 0.988 | |
| RFHardware=Nexus 5, Layer Type=Convolutional2018.09 | 19.7 | 27.3 | 0.985 | |
| RFDescription=random forest regression, Input features=explanatory variable vector x, Device=Nexus 5, Target layer=Convolutional layer2018.09 | 19.7 | 27.3 | 0.985 | |
| DTHardware=Nexus 5, Layer Type=Convolutional2018.09 | 23.8 | 39.2 | 0.969 | |
| DTDescription=classification and regression trees, Input features=explanatory variable vector x, Device=Nexus 5, Target layer=Convolutional layer2018.09 | 23.8 | 39.2 | 0.969 | |
| SVRHardware=Nexus 5, Layer Type=Convolutional2018.09 | 233.8 | 227.1 | -0.229 | |
| SVRKernel=radial basis function, Input features=explanatory variable vector x, Device=Nexus 5, Target layer=Convolutional layer2018.09 | 233.8 | 227.1 | -0.229 |