What's an artifact store and why separate from model registry?

medium

Answer

$Artifact store: raw binary files (weights, tokenizers, preprocessors, plots, data snapshots).$
$Model registry: metadata layer above artifacts with lifecycle (staging \to prod).$
$Analogy: artifacts = files, registry = database of blessed versions.$
$Separation: registry versions can point to artifacts anywhere (S3, GCS).$
$Enables: (1) storage-agnostic registry.$
$(2) same artifact referenced by multiple registry entries.$
$(3) retention policies at different layers.$
$MLflow: runs:/id/model (artifact) vs models:/name/version (registry).$

Check yourself — multiple choice

Same
Artifact store = raw binaries (weights/tokenizer/plots); model registry = metadata + lifecycle (staging→prod); artifacts=files, registry=DB of blessed versions; MLflow runs:/vs models:/
Random
Not real

Artifact store = files; registry = metadata + lifecycle.

#reproducibility#mlops

215 interview questions in this topic.