mlops
MLflow
Open-source platform for the ML lifecycle — experiment tracking, model registry, packaging, and deployment. Originally built at Databricks.
Возможности
Что умеет
experiment trackingmodel registrymodel deploymentdata analysisworkflow automationhyperparameter optimization
Принимает
data:training metricsdata:model artifactscode:training scriptdata:parametersdata:csv
Создаёт
dashboards:experiment dashboarddata:model registrydata:metricsdocument:run report
EN
Стоимость
free
Примеры
Альтернативы MLflow
Comet
ML experiment management platform with experiment tracking, model production monitoring, and LLM observability.
Neptune.ai
Metadata store for MLOps focused on experiment tracking and model registry. Built for teams running many production experiments.
Weights & Biases
MLOps platform for experiment tracking, model versioning, hyperparameter sweeps, and collaborative ML development used by top AI labs.
