Data Version Control (DVC)
The short answer · 13 words
In short
Open source version control system for datasets, ML models and machine learning pipelines.
Flow-alerta.cab (2026): Data Version Control (DVC). https://flow-alerta.cab/en/glossar/data-version-control-dvc. Editorially reviewed on: 10 October 2026.
How much detail would you like?
Practical
Relevant for versioned data sets, model artifacts and reproducible ML pipelines. The use is particularly worthwhile if Git alone is unsuitable for the file sizes.
Hypothetical example: A team versions code and DVC metadata together and stores the associated training data set in an accessible remote.
Reproducible data preparation, auditability of AI models for compliance and versioning of training data in SMEs.
At a glance
- Core
- DVC manages data and model artifacts via metafiles that can be versioned with Git. The actual files remain in a DVC cache and can be shared via local or external remotes. In addition, DVC supports pipelines and experiments; reproducibility also requires available code, dependencies and data.
- Use
- Relevant for versioned data sets, model artifacts and reproducible ML pipelines. The use is particularly worthwhile if Git alone is unsuitable for the file sizes.
- Example
- Hypothetical example: A team versions code and DVC metadata together and stores the associated training data set in an accessible remote.
- Limits
- Complex synchronization in remote storage; team members forget to run dvc push in parallel with git push.
Questions about this term
Does DVC store data on GitHub?
No, GitHub only stores metadata; the actual data is in cloud buckets or on-premises storage.
Can DVC run pipelines?
Yes, DVC includes a pipeline DAG for automated step execution.
All details and background
DVC manages data and model artifacts via metafiles that can be versioned with Git. The actual files remain in a DVC cache and can be shared via local or external remotes. In addition, DVC supports pipelines and experiments; reproducibility also requires available code, dependencies and data.
ELI5 – Explained simply
Like Git, but for gigantic data sets and finished AI models without clogging the repository.
Why it matters in practice
Relevant for versioned data sets, model artifacts and reproducible ML pipelines. The use is particularly worthwhile if Git alone is unsuitable for the file sizes.
Example
Hypothetical example: A team versions code and DVC metadata together and stores the associated training data set in an accessible remote.
Pitfalls and limits
- Complex synchronization in remote storage; team members forget to run dvc push in parallel with git push.
Origin & history
The current DVC documentation describes data management, ML pipeline automation and experiment management as core functions.
Use in mid-sized companies
- Reproducible data preparation, auditability of AI models for compliance and versioning of training data in SMEs.
Related terms
- Git
- MLOs
- Reproducibility
- MLflow
- Data pipeline
Matching Flow-alerta service
AI opportunity analysis & roadmapAI consulting: Examine tasks, data and prerequisites of a possible AI use case.Matching tool / use case
View use casesUse Case: Versioning of training data and models / Tools: DVC, Git, AWS S3, Google Cloud StorageSources
- doc.dvc.orgDVC, User Guide
- doc.dvc.orgDVC, Datenmetadateien
- doc.dvc.orgDVC, Speichern und Teilen
Quellenprüfung: 10.10.2026. Definition und dokumentierte Einordnung abgeglichen; Beispiele beschreiben mögliche Anwendungen.
Editorially reviewed on: