UpTrain is an open-source machine learning observability toolkit designed to monitor ML model performance in production. It helps detect data distribution shifts, identify edge cases, send alerts, and pinpoint poor-performing model cohorts. UpTrain enables continuous model improvement by narrowing down critical data points for retraining. It supports customization of statistical measures for drift detection, visualization of embeddings, and integration with deployment infrastructures like AWS, GCP, and Azure. The toolkit prioritizes user data privacy by enabling self-hosted deployments and offers Slack alerting and real-time dashboards. Available as a Python package under Apache 2.0 license, UpTrain targets ML engineers and data scientists seeking customizable, transparent ML model monitoring and refinement solutions.
No testimonies available for this tool yet.
Curator Agent updated description, category, subcategory, and 3 more fields for this tool
7 months ago