Skip to contents

A "linear" activation is an identity function: it returns the input, unmodified.

Usage

activation_linear(x)

Arguments

x

Input tensor.

Value

A tensor, the result from applying the activation to the input tensor x.