Intel® Geti™ 1.5.1#

Release Summary#

The Intel® ™ 1.5.1 release is now available; the updates and fixes include the following:

  • New storage usage bar display – Enables monitoring disk usage with a visual indicator.

  • Functional alert for low disk space - For optimal performance, the platform will now alert users to clear disk space when it goes below a threshold.

  • Resolved repeated label issue with the stamp tool in task chains - The stamp tool no longer adds an extra detection label when copying annotations from one frame to another.

  • Uninterrupted import of COCO datasets - A 504 Gateway Timeout issue during COCO dataset import has been resolved.

  • Remove child labels without disrupting model training - Users can now delete child labels in hierarchical classification tasks without leading to failure in training tasks.

  • Corrected model size in the download column. - Obtain accurate model size information in the Models tab.

  • Fixed model creation date in breadcrumb - Now, users can accurately track the chronology of their models.

Release Details#

This section covers additional details on the updates and fixes available with Intel® ™ 1.5.1.

New storage usage bar display#

To assist users in managing their disk space, a storage usage bar is now available on the Workspace screen. This visual indicator shows disk space usage, allowing users to manage storage more effectively and efficiently.

Placeholder

Functional alert for low disk space#

A warning message will now appear when the free storage limit reaches 20 GB to prevent the platform from becoming unresponsive due to limited storage. When 15 GB space or less is reached, the platform will disable the import dataset, project upload, and training functions. When 5 GB or less is reached, the platform stops working. This early alert system allows users to clear disk space and maintain platform performance.

Resolved label issue with the stamp tool in task chains#

The misbehavior in stamp tool that created extra detection labels when reusing the annotation in task chains has been resolved. Users can now seamlessly annotate across frames and images without adding extra labels.

Uninterrupted import of COCO datasets#

A 504 Gateway Timeout error that occurred while importing COCO datasets has been resolved. Users can now import these datasets without interruptions.

Remove child labels without disrupting model training#

An issue where training would fail after a child label was deleted in classification tasks has been resolved. Users can now delete child labels without disrupting the training process.

Corrected model size in download column#

Obtain accurate model size information in the Models tab.

Fixed model creation date in breadcrumb#

Now, users can accurately track the chronology of their models.