MLflow

MLflow

MLflow is an open-source platform for managing the complete machine learning lifecycle. Created at Databricks and now an Apache project, it covers experiment tracking, reproducible runs, a central model registry, and one-click model deployment — with first-class support for LLM evaluation and prompt engineering workflows.

Key Features

• Experiment tracking: log parameters, metrics, artifacts, and code version per run
• Model Registry: versioned model store with staging/production promotion
• MLflow Tracking UI: compare runs with interactive plots
• LLM evaluation: score prompts/chains with built-in metrics (toxicity, faithfulness, BLEU, ROUGE)
• Model serving: REST API server for any logged model
• Integrates with scikit-learn, PyTorch, TensorFlow, Keras, XGBoost, LightGBM, Spark, LangChain

Quick Start

import mlflow

with mlflow.start_run():
    mlflow.log_param("lr", 0.01)
    mlflow.log_metric("accuracy", 0.92)
    mlflow.sklearn.log_model(model, "model")

# Launch UI
# mlflow ui

Install via ai-supply

npx ai-supply add mlflow-experiment-tracking

Curated mirror of the open-source MLflow (Apache-2.0). Get it from the source.