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

Defines pruning and quantization with tradeoffs and an API call, but does not detail pruning algorithms or evaluation procedures.

Model optimization

As explained in Deep Learning with R, Third Edition

Model optimization is the modification of a trained model to reduce computational, memory, or latency costs while preserving acceptable predictive performance.

Model optimization modifies a trained model to reduce inference memory, computation, or latency while preserving acceptable predictive performance. It matters especially on smartphones, embedded devices, and other constrained or low-latency targets. section 6.3.2

Use two main techniques to reduce inference cost:

Apply weight pruning to retain the most significant coefficients, reducing parameters and computation while exposing a tunable size-versus-accuracy tradeoff. section 6.3.2
Apply weight quantization to convert float32 weights to int8, producing an inference-only model that is four times smaller and near the original accuracy. section 6.3.2

Quantizing a Keras model

Call quantize_weights(model, "int8") before importing the model into TensorFlow.js or exporting it to TensorFlow Lite. Validate the resulting accuracy because optimization trades resource usage against predictive performance. section 6.3.2

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

Machine Learning System DesignModel optimization is the improvement of a trained model's execution efficiency or resource requirements while preserving acceptable predictive behavior.

Compare how 2 books explain model optimization →

Related concepts

pruningincludes · Pruning is the selective removal or restriction of parts of a model, structure, or candidate set, typically to reduce complexity or computation while retaining useful results.
deploymentconfigures · Deployment is the process of making software, models, or other computational systems available to operate in a target environment.
quantizationincludes · Quantization maps numerical values to a finite set of discrete representations, often using reduced precision to lower model storage or computation requirements.
distillationincludes · Knowledge distillation is a training technique in which a student model learns to reproduce or approximate the outputs or behavior of a teacher model.
ONNXaddressed by · ONNX is an open standard for representing machine-learning models in a portable format that supports exchange and execution across frameworks and runtimes.

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.