MLflow is a package for the Datagrok platform.
MLflow integrates MLflow models into the Datagrok platform. It lets users fetch, manage, and run MLflow-registered models within Datagrok’s scalable infrastructure using Docker containers.
Key features:
- Model fetching: Easily access MLflow models using the MLflow connector
- Containerized execution: Deploy models in isolated Docker containers for secure and reproducible execution
- Inference API: Perform real-time predictions and visualize results within Datagrok’s interactive environment
- Version control: Access specific or latest versions of models with MLflow’s model versioning
- Environment consistency: Maintain consistent environments using MLflow’s conda or Docker specifications
- Seamless integration: Combine Datagrok’s analytics and visualization with ML model inference
Use cases:
- Perform automated model inference on large datasets within Datagrok
- Collaborate on data science workflows integrating data wrangling, visualization, and machine learning