"Class" (Net Encoder)

[Experimental]

NetEncoder[{"Class",classes}]

represents an encoder that converts a class label in the list classes to an integer code.

NetEncoder[{"Class",classes,form}]

represents an encoder that converts class labels to the output type form.

Details

NetEncoder[…][input] applies the encoder to an input to produce an output.
NetEncoder[…][{input₁,input₂,…}] applies the encoder to a list of inputs to produce a list of outputs.
The input to the encoder input_i must be an element in the list classes.
The following parameters are supported:
"Dimensions" {} input dimensions

"Multilabel" False whether classes are not exclusive
When "Multilabel"False and form is "Index" (the default), the output of the encoder is the integer that is the position of input in classes.
When "Multilabel"False and form is "IndicatorVector", the output of the encoder is an n-dimensional unit vector in the p direction, where p is the position of input in classes, and n is the length of classes.
When "Multilabel"True and form is "IndicatorVector", the output of the encoder is an n-dimensional unit vector with ones at the position of input classes and zeros elsewhere.
An encoder can be attached to an input port of a net by specifying "port"->[…] when constructing the net.

Examples

open allclose all

Basic Examples (1)Summary of the most common use cases

Create a class encoder with two classes:

In[1]:=1

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-v9nmdh

Out[1]=1

Apply it to a batch of inputs:

In[2]:=2

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-8ku6q5

Out[2]=2

Scope (1)Survey of the scope of standard use cases

The default output form is to encode classes as integers:

In[1]:=1

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-bddxmk

Out[1]=1

In[2]:=2

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-3v49em

Out[2]=2

Specifying the "IndicatorVector" form will encode classes as "one-hot" vectors:

In[3]:=3

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-hvyfbr

Out[3]=3

In[4]:=4

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-cp4tds

Out[4]=4

Parameters (3)

"Dimensions" (2)

Create an encoder that takes a 3x3 matrix of classes and returns a matrix of indices:

In[1]:=1

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-pdsdod

Out[1]=1

Apply it to a single input:

In[2]:=2

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-ln4n2t

Create an encoder that takes a 2x2 matrix of classes and returns unit vectors:

In[1]:=1

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-s85abb

Out[1]=1

Apply it to a single input:

In[2]:=2

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-5kg3fa

Out[2]=2

"Multilabel" (1)

Create a multilabel encoder:

In[1]:=1

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-t284xu

Out[1]=1

Apply it to a single input with two classes:

In[2]:=2

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-sqiqm6

Out[2]=2

Apply it to a batch of inputs:

In[3]:=3

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-sp20gy

Properties & Relations (1)Properties of the function, and connections to other functions

NetTrain will automatically try attach an encoder when a net is not fully specified. Automatic attachment of a class encoder:

In[1]:=1

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-on9q0n

Out[1]=1

In[2]:=2

✖

https://wolfram.com/xid/0jvxou418hvrde4a2d2q-cma7rq

Out[2]=2

Top