Implements Dot Product Attention using the given inputs.
An ElementWiseVertex is used to combine the activations of two or more layer in an element-wise manner
For example, the activations may be combined by addition, subtraction, multiplication (product), average or by selecting the maximum.
FrozenVertex is used for the purposes of transfer learning.
A GraphVertex is a vertex in the computation graph type of neural network.
L2NormalizeVertex performs L2 normalization on a single input, along the specified dimensions.
L2Vertex calculates the L2 (Euclidean) least squares error of two inputs, on a per-example basis.
LayerVertex is a GraphVertex with a neural network Layer (and, optionally an
A MergeVertex is used to combine the activations of two or more layers/GraphVertex by means of concatenation/merging.
Removes the first column and row from an input.
PreprocessorVertex is a simple adaptor class that allows a
Adds the ability to reshape and flatten the tensor in the computation graph.
A ScaleVertex is used to scale the size of activations of a single layer: this is simply multiplication by a fixed scalar value
For example, ResNet activations can be scaled in repeating blocks to keep variance under control.
A ShiftVertex is used to shift the activations of a single layer.
StackVertex allows for stacking of inputs so that they may be forwarded through a network.
SubsetVertex is used to select a subset of the activations out of another GraphVertex.
UnstackVertex allows for unstacking of inputs so that they may be forwarded through a network.
Copyright © 2020. All rights reserved.