Value Members

1.  object ConvNetwork extends Serializable
   Since

   31.08.17

2.  object DefaultNetwork extends Serializable

   This is a feed-forward neural network with fully connected layers.

   This is a feed-forward neural network with fully connected layers. It uses dynamic gradient descent to optimize the error function Σ1/2(y - net(x))².

   Use the parallelism parameter with care, as it greatly affects memory usage.

   Since

   15.01.16

3.  object LBFGSNetwork extends Serializable

   This is a fully connected Neural Network that uses Breeze's LBFGS, a quasi-Newton method to find optimal weights.

   This is a fully connected Neural Network that uses Breeze's LBFGS, a quasi-Newton method to find optimal weights.

   Since

   12.06.16

4.  object LSTMNetwork extends Serializable

   This is a Long-Short-Term Memory Network.

   This is a Long-Short-Term Memory Network. The standard model is implemented. It comes with recurrent connections and a dedicated memory cell with input-, output- and forget-gates for each neuron. Multiple layers can be stacked horizontally, where the current layer gets input from the lower layers at the same time step and from itself at the previous time step.

   Since

   07.07.16