Annotation Lab has improved its support for Visual NER projects. Visual NER pre-trained models are now available on the NLP Models Hub and can be added to your configuration using the Visual UI. Page navigation on multipage PDF tasks is improved and tasks can be loaded directly from s3.
Visual NER Models available in the Models Hub page
Visual NER models can now be filtered, downloaded from the NLP Models Hub and used for pre-annotating image-based documents.
Once you download the models from the Models Hub page, you can see the model’s label in the Predefined Label tab on the project configuration page.
Visual configuration options for Visual NER project.
Users are now able to add custom labels and choices in the project configuration from the Visual tab for Visual NER projects as well as for the text projects.
Improved page navigation for Visual NER projects
For Visual NER projects, users can jump to a specific page in any multi-page task, instead of passing through all pages to reach a target section of a PDF document.
Import tasks from s3
NLP Annotation Lab 4.3.0 offers support for importing tasks/documents stored on Amazon S3. In the `Import Page`, a new section was added which allows users to define S3 connection details (credentials, access keys, and S3 bucket path). All documents present in the specified path, are imported as tasks in the current Annotation Lab project.
Project-level history of the Trained Models
It is now possible to keep track of all previous training activities executed for a project. When pressing the `History` button from the `Train` page, users are presented with a list of all trainings triggered for the current project. Each training event is characterized by the source (manual, active learning), data used for training, date of event, and status. Training logs can be downloaded for each training event.
Easier page navigation
Since version 4.0.0, users were not able to right-click on the available links and select “Open in new tab”. This feature has been added so that users can easily open any link in a new tab without losing the current work context.
Optimized user editing UI
The user add/edit form was optimized in this version. All the checkboxes on the Users Edit page now have the same style. The “UserAdmins” group was renamed to “Admins” and the description of groups is more detailed and easier to understand. Also, a new error message shown when an invalid email address is used was updated.
Stay tuned for more exciting news!