Bayesian Optimization on CNN hyperparameter tuning
6Mean Stopping IterationOracler
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OraclerKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 6 | 100 | 0.0208 | |
| ∆ESKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 6.5 | 100 | 0.0207 | |
| AcqKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 7 | 100 | 0.0205 | |
| PRBKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 10.9 | 100 | 0.0204 | |
| ∆CBKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 11 | 100 | 0.0204 | |
| UCBbrKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 34.3 | 100 | 0.0177 | |
| NOSTOPKernel=SE kernel, Dimension (D)=4, Budget (T)=2562026.05 | 256 | 100 | 0.0148 |