You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
When using unlabelled annotations for few-shot learning, the response extractor has trouble extracting the responses because a map with null keys is not supported. Also sampling of examples for the few-shot learning can be improved to cover the set of labels better and to prefer having multiple example mentions per example.
The text was updated successfully, but these errors were encountered:
* main:
#4638 - Ability to filter annotation sidebar by covered text
#4636 - Ability to collapse groups in annotation sidebar
#4634 - Pull document file storage code out of document service into a new service
#4550 - Clean up code
#4625 - Azure AI Open AI recommender
#4628 - Better results when using unlabelled annotations for few-shot learning
#4550 - Clean up code
#4625 - Azure AI Open AI recommender
* main:
#1321 - Importing .ann files from BRAT
#4614 - Display suggestions from Named Entity Linker immediately
#4638 - Ability to filter annotation sidebar by covered text
#4636 - Ability to collapse groups in annotation sidebar
#4634 - Pull document file storage code out of document service into a new service
#4550 - Clean up code
#4625 - Azure AI Open AI recommender
#4628 - Better results when using unlabelled annotations for few-shot learning
#4550 - Clean up code
#4625 - Azure AI Open AI recommender
#1321 - Importing .ann files from BRAT
[maven-release-plugin] prepare for next development iteration
[maven-release-plugin] prepare release inception-31.3
#1321 - Importing .ann files from BRAT
#4622 - No helpful error message when recommender is not called due to feature being disabled
No issue: Better logging
#4550 - Clean up code
Is your feature request related to a problem? Please describe.
When using unlabelled annotations for few-shot learning, the response extractor has trouble extracting the responses because a map with null keys is not supported. Also sampling of examples for the few-shot learning can be improved to cover the set of labels better and to prefer having multiple example mentions per example.
The text was updated successfully, but these errors were encountered: