Artifact packaging

As explained in Machine Learning System Design

Artifact packaging is the process of bundling a trained model and its supporting files into a deployable format.

Artifact packaging bundles a trained model and its supporting files into a deployable format. In a typical training pipeline, it follows data fetching, preprocessing, model training, evaluation and testing, postprocessing, and report generation. section 10.1.1

Use artifact packaging as the final checkpoint in a training-pipeline checklist:

• Complete model evaluation and testing before packaging the trained model. section 10.1.1
• Keep the training pipeline distinct from the lower-level inference pipeline, while recognizing that training can contain inference steps. section 10.1.1

Figure 10.1 A DAG scheme representing a training pipeline

Use the training-pipeline DAG to locate artifact packaging among the pipeline’s stages. section 10.1.1

How Machine Learning System Design covers it

How deep it goes

Identifies artifact packaging as a pipeline step but does not detail its files, format, or implementation.

Examples

Uses a typical training pipeline’s final artifact-packaging step as its concrete example.

What it assumes

Requires distinguishing training pipelines from lower-level inference pipelines.

What you can do after

Apply it by placing artifact packaging after report generation in a training-pipeline checklist.

How it connects

Connects artifact packaging with training pipelines, preprocessing, model training, model evaluation, postprocessing, and report generation.

liveBookmanning.com

Read Machine Learning System Design on liveBook Get the book at manning.com

Related concepts

• ONNXuses · ONNX is an open standard for representing machine-learning models in a portable format that supports exchange and execution across frameworks and runtimes.
• containerizationuses · The packaging of software and its dependencies into isolated, portable execution environments called containers.
• pipelinepart of · A pipeline is an ordered or coordinated sequence of connected processing stages or tasks that passes data, work, or results toward an outcome.

This explainer was generated by Manning's knowledge graph from the text of Machine Learning System Design. The section links point to the parts of the book it draws on.