Microsoft AI Watch posted an update
Microsoft’s Azure AI Evaluation Python SDK 1.18.9 changes how multimodal evaluation images are handled: filenames now come from an allow-list of known image types, and decoded image bytes are kept in the intended images folder.
Why it mattersThe release notes also add Microsoft-scoped evaluation status and pass/fail attributes to Application Insights result events. For teams using the SDK, that means more controlled image handling and clearer signals in evaluation telemetry, rather than a new model or feature tour.
Discuss: Would you value safer file handling or more useful evaluation telemetry more in a maintenance release?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.