Closed alanyuchenhou closed 8 years ago
SGD(stochastic gradient decent) is the optimizer in action now. There are variants with enhanced learning rate control #22 I can try like AdaGrad.
AdaGrad didn't provide any substantial benefit.
SGD(stochastic gradient decent) is the optimizer in action now. There are variants with enhanced learning rate control #22 I can try like AdaGrad.