- Add support for training/holdout dataset assignment as model version level
- Internal improvements
- Bug fixes
- Add support for LLM (text generation) custom models
- Add support for egress network policy for a custom model
- Bug fix
- Bug fix
- Bug fixes
- Enable users to specify a reason for a model replacement in a deployment
- Separately track model's settings from its versions
- Fetch and print deployment's log in case of a deployment's creation failure
- Add a missing mapping of Anomaly target type
- Internal improvements
- Bug fixes
- Improve segment analysis functionality and update README.md
- Bug fix
- Fix examples in the README.md to comply with the defaults in DataRobot
- Internal improvement
- Add support for namespaces
- Add a new metric to count models and deployment whose settings were changed
- Internal improvements
- Internal improvements
- Bug fixes
- Update README.md
- Internal improvements for local utility
- Bug fixes
- Change attribute names in a deployment schema:
prediction_id
==>association_id_column
actuals_id
==>actual_values_column
- Update README.md regarding
importance
attribute in a deployment. - Update README.md regarding the user's provided ID.
- Bug fixes.
- Improve documentation in the README.md file
- Internal improvements to the development environment
- Bug fixes
- Add a 'Quick Start' section to the README.md.
- Update the action's name and add branding section.
- Update the README.md about referencing specific releases.
- Add support for custom model versions' dependencies.
- Add a script to create a release.
- Bug fixes
- Bug fix
- Update the way statistics are emitted by the action.
- Functional tests are executed against DataRobot US production system.
- The first release.