How deeply does Deep Learning with R, Third Edition cover variable?

Defines creation, full and indexed assignment, additive updates, and gradient tracking, but does not develop optimizer internals.

Variable

As explained in Deep Learning with R, Third Edition

A TensorFlow variable is a mutable tensor-like object that stores modifiable model state and can be updated by assignment.

A TensorFlow variable is a mutable tensor-like object for storing model state that changes during computation. Unlike a regular tensor, it supports assignment and in-place-style updates. section 3.3.1 section 3.2

Create and update a variable

Create a variable from a zero-filled tensor, replace its contents, then add to every element:\n\n```r\nv <- tf$Variable(initial_value = tf$zeros(shape = shape(3, 1)))\nv$assign(tf$ones(shape(3, 1)))\nv$assign_add(tf$ones(shape(3, 1)))\n```\n\nAssignment changes the stored value, whereas direct item assignment on an ordinary tensor is unsupported. section 3.3.1

Variables also participate naturally in automatic differentiation: GradientTape tracks operations involving them and can compute gradients with respect to their values. section 3.2

How Deep Learning with R, Third Edition covers it

Cover of Deep Learning with R, Third Edition

liveBook manning.com

Read Deep Learning with R, Third Edition on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Deep Learning with Python, Third EditionA mutable TensorFlow value that stores learnable parameters such as weights and bias in this example.

Compare how 2 books explain variable →

Related concepts

model parallelismused by · Model parallelism is a distributed-computing technique that partitions a machine-learning model’s parameters, layers, or computations across multiple devices.
tensoruses · A tensor is a multidimensional, array-like structure of values indexed along zero or more dimensions and used for numerical computation.
mathematical variableimplements · A mathematical variable is a symbol or quantity representing a value or attribute that can vary or be determined within a defined domain.
variablekind of · A variable is a named program binding or entity associated with a value that can be accessed and, when permitted, changed during computation.

This explainer was generated by Manning's knowledge graph from the text of Deep Learning with R, Third Edition. The section links point to the parts of the book it draws on.