V11.1.1¶
August 13, 2025
The DataRobot Self-Managed AI Platform v11.1.1 release, part of DataRobot's continual effort to secure the product and platform, addresses issues reported in earlier releases. See also the features introduced in v11.1.0, released July 17, 2025.
Issues fixed in V11.1.1¶
The following issues have been fixed since release v11.1.0:
Generative AI¶
- BUZZOK-27070: Fixes an issue where the LLMs dropdown list that appeared when creating a new LLM blueprint in a playground would error due to all external LLM provider access/feature flags having been disabled.
- BUZZOK-26971: Fixes an issue with potential logging of prompts/responses in an agentic playground in the case of unhandled exceptions inside of agent code.
- BUZZOK-26870: Fixes the Faithfulness and Correctness metrics computation in the playground UI.
- BUZZOK-27088: Updates dependencies of the "[DataRobot] Python 3.11 GenAI" global execution environment to ensure global tools are installable offline in Self-Managed deployments.
Modeling¶
- MODEL-20123: Allows users to set the memory sizes of a workers based on the dataset size for all data chunking jobs. This is limited to Self-Managed deployments.
- MODEL-19906: Fixes an issue where the SHAP matrix insight calculation would fail if geolocations contained missing values.
- MODEL-19731: Fixes an issue where the Feature Effects job would fail to start if called from a "SHAP-only mode" project and the prerequisite SHAP Impact job had been submitted using the
datarobot.insights.ShapImpactmethod from the Python API client. Now, the prerequisite job is found successfully whether submitted by that method, the legacy methoddatarobot.ShapImpact, or by the UI. - MODEL-18975: Improves handling of PDF image parsing and runtime errors (e.g., due to no supported image decoder) in the OCR-service. When an input PDF contains unsupported images, the error message will be saved to the error report. The PDF input will be returned as-is.
Model management¶
- MMM-19984: Fixes an issue where a model couldn't be deployed because the XEMP Prediction Explanations initialization job failed. Now, by default, the prepare-for-deployment workflow uses a much less resource intensive version of the XEMP init job.
- MMM-19955: Fixes an issue where boolean (True/False) values were not rendering correctly in the tracing chart details panel.
Notebooks¶
- CFX-3290: Fixes an issue where if a notebook is converted to a codespace, the notebook's content did not appear in the newly created codespace.
Platform¶
- PLT-19024: Fixes user membership pages for user accounts with defined seat licenses.
All product and company names are trademarks™ or registered® trademarks of their respective holders. Use of them does not imply any affiliation with or endorsement by them.