Why MLEM_
- Unambiguous link between data, code, model and metrics
- Standardised model packaging with environment and input data specification
- Model lifecycle management using GitOps approach
- Cloud-agnostic model deployment
Open-source model registry and deployment tool for Machine Learning_
Codify your model into a standardized format
Automatically extract environment, methods, and input data specification
Turn your Git repo into model registry
Reuse existing Git and GitHub/GitLab infrastructure for model management
Use CLI to pack, dockerize and deploy
Easily switch between different packaging formats and cloud providers
Use Python API to load and apply your models
Load models dynamically from any storage or model registry