Deepchecks is an open-source tool designed for the continuous validation of machine learning (ML) models and data. It provides a comprehensive solution for testing and monitoring AI and ML models from research stages through to production. The platform includes built-in and customizable checks for various data types, including tabular data, natural language processing (NLP), and computer vision (CV). These checks help identify issues related to model performance, data distribution, and data integrity.
Deepchecks offers several components:
-
Deepchecks Testing: This component allows users to run both built-in and custom checks and suites for validating tabular, NLP, and CV data. It supports collaboration over test results and efficient iteration until the model is production-ready.
-
CI & Testing Management: This managed offering facilitates collaboration over test results and efficient iteration until the model can be deployed. It is currently in closed preview.
-
Deepchecks Monitoring: This component tracks and validates the behavior of deployed models in production, allowing for continuous monitoring and alerting.
The tool supports integration with CI/CD pipelines and can be installed via pip or conda. It provides visual reports and JSON outputs to inspect the results of checks and suites. Additionally, Deepchecks includes a dynamic UI for examining test results and monitoring production models.
Deepchecks is primarily open-source, released under the AGPL 3.0 license, with some premium features available under a commercial license. The community-driven project encourages contributions and offers extensive documentation, tutorials, and support through Slack and GitHub.
Pricing
Pricing information is not available