You can now stop a running ML Skill manually. You can do this if you want better manage licenses but at the same time keep the slot for the skill.
For more information on ML Skills, check the Managing ML Skills page.
Since we are now allowing you to download ML Package directly, we are not exposing the zip file as a pipeline artifacts anymore. If you want to download it, just download the ML Package version listed as Generated version on Pipeline details page.
- Fixed an issue that occasionally caused a
Connection refusederror while using ML Skills.
- Fixed an issue where searching for users using the people-picker was not working properly.
There are now two new fields when viewing a Pipeline information:
- Generated version: the version generated by the pipeline.
- Base version: the initial version.
You can use this information to quickly find out which version was generated by a pipeline.
- The default order for Out of the Box Packages has changed. The packages are now ordered by their version number so it's easier to pick the latest version.
- Fixed an issue where the Import ML Package button on the ML Packages page was managed by the wrong permission on project level.
- Fixed an issue where the Delete user permission was not working properly.
Back in May we announced new project-level and tenant-level access control for new tenants.
We are excited to release a new procedure on how to update your existing tenants to the new access control.
Users with all old permission (ML Skills, ML Packages and ML Logs) will be granted the Administrator role for AI Center after the update.
To update your tenant to the new access control configuration, follow the steps below.
- Click on the three dot icon in the AI Center homepage.
- Select Update to new access controls from the drop-down list.
- Click Update.
There is no downtime expected during update.
All projects will have restricted access and existing users will automatically be added to the project after the update.
For more information on this, check our forum.
We are happy to announce that we introduced new features regarding AI Units in the case of Pipelines and ML Skills.
You can check the AI Units consumption when deploying a Pipeline or ML Skill.
For ML Skills deployment, you can now configure advanced infra settings. Boy doing this, you can process more pages in parallel if the model is called several times in parallel. This is optional and only recommended for advanced users.
For more information on how to do this, check the Managing Pipelines and Managing ML Skills pages.
Note: AI Units are deployed progressively. Check our release notes for future information.
- The telemetry process is updated, improving the process of collecting and transmitting the needed information.
- The people picker for assigning people to AI Center or to specific project is improved and the interface is updated.
- UiPath AI CenterTM is dependent on UiPath Document Understanding.
You can have Document Understanding enabled without AI Center, but make sure to enable Document Understanding product on every tenant where you need to use AI Center.
- UiPath Task Mining is dependent on AI Center, which is dependent on Document Understanding.
If you want to enable Task Mining, AI Center is enabled automatically but you need to manually enable Document Understanding.
The date when a change is first announced in the release notes is the date when it first becomes available.
If you don't see the change yet, you can expect to see it soon, after we roll out changes to all the regions.
Updated about a month ago