MLflow Skill

Expert system for MLflow operations, tracking, and model management.

When to Use This Skill

This skill should be triggered when:

•Working with MLflow Tracking, Projects, Models, or Registry
•Looking up MLflow API references or feature documentation
•Implementing MLflow integration in machine learning pipelines
•Debugging MLflow-related errors or issues
•Seeking best practices for experiment tracking and model deployment

Quick Reference

Common Patterns

Pattern 1: Install MLflow

bash

pip install mlflow[auth]

Pattern 2: Command Line Help

bash

mlflow --help

Pattern 3: Install Extras

bash

# Time series
pip install mlflow[sktime]
pip install mlflavors
# Visualization
pip install mlflow-vizmod
# Cloud
pip install bigmlflow
pip install mlflow[aliyun-oss]

Pattern 4: Start UI

bash

mlflow ui

Pattern 5: referencing models

code

models:/<model_name>/<version>
models:/<model_name>/Production

Pattern 6: Log Model (PyTorch example)

python

mlflow.pytorch.log_model(model, "model")

Example Code Patterns

Example 1: Configure S3 Endpoint (Bash)

bash

export MLFLOW_S3_ENDPOINT_URL=https://s3.us-west-1.amazonaws.com

Example 2: Set Bucket Owner (Bash)

bash

export MLFLOW_S3_EXPECTED_BUCKET_OWNER=123456789012

Reference Files

This skill includes comprehensive documentation in references/:

•ml.md - Complete MLflow documentation covering all major components.

Use the view_file tool to read specific sections of these files when the user needs detailed explanations.

usage

For Beginners

Start by reading the overview or getting started sections in references/ml.md to understand the core concepts of Runs, Experiments, and the Tracking Server.

For API Lookups

Search for the specific function or class name (e.g., mlflow.log_param, mlflow.sklearn.log_model) within references/ml.md to find parameters and usage examples.

For Troubleshooting

Check the reference documentation for configuration options and common pitfalls, especially regarding authentication and remote server connectivity.