models

Type Members

1. trait Model extends Logging

   

   Abstract base class for neural net architectures.

2. class NeuralNet extends Model with Logging

   

   Simple DL4J-style sequential neural net architecture with one input node and one output node for each node in computational graph.

   Simple DL4J-style sequential neural net architecture with one input node and one output node for each node in computational graph.

   Wraps DL4J MultiLayerNetwork. Enforces DL4J model construction pattern: adds pre-processing layers automatically but requires user to specify output layer explicitly.

3. class Sequential extends Model with Logging

   

   Class for keras-style simple sequential neural net architectures with one input node and one output node for each node in computational graph.

   Class for keras-style simple sequential neural net architectures with one input node and one output node for each node in computational graph.

   Wraps DL4J MultiLayerNetwork. Enforces keras model construction pattern: preprocessing (reshaping) layers should be explicitly provided by the user, while last layer is treated implicitly as an output layer.