net

Type Members

1. class GraphNet[T] extends Container[Activity, Activity, T] with NetUtils[T, GraphNet[T]] with Predictable[T]

   

2. class IdentityCriterion extends AbstractCriterion[Activity, Activity, Float]

   

3. class MergeFeatureLabel extends ImageProcessing

   

4. trait NetUtils[T, D <: Module[T] with NetUtils[T, D]] extends AnyRef

   

5. class TFNet extends AbstractModule[Activity, Activity, Float] with Predictable[Float]

   

   TFNet wraps a tensorflow subgraph as a layer, and use tensorflow to calculate the layer's output.

   TFNet wraps a tensorflow subgraph as a layer, and use tensorflow to calculate the layer's output.

   This subgraph should not contain any tensorflow Variable and the input/output must be numeric types

   When used with other layers for training, there should be no trainable layer before this one, as the gradInput of this layer is always zero.

6. class TFOptimizer extends AnyRef

   

7. class TFValidationMethod extends ValidationMethod[Float]

   

8. case class TrainMeta(inputNames: Array[String], outputNames: Array[String], variables: Array[String], gradVariables: Array[String], defaultTensorValues: Array[Array[Float]]) extends Product with Serializable