Annotate a list of user utterances instead of one at a time in Rasa X

(Dinesh) #1

In Rasa X, there is a functionality to “Annotate new data” (Training > NLU Training > Annotate new data > ‘+’) as shown in the picture below. But, it allows only one sentence (or user utterance) to be classified for the intent at a time. Is there a way/workaround to do the same, but for a list of sentences (say a text file containing a list of sentences separated by new line character)? This would greatly expedite the intent annotation process, especially in the projects that use only the rasa nlu.

(Ella Rohm-Ensing) #2

Hey there @enigmane, is your issue that you want to upload a bunch of training data that’s already grouped into intents? If so you can do it here

with the up arrow button.

(Dinesh) #3

No. That is for the case of “a bunch of training data that’s already grouped into intents”. What if a bunch of unlabeled utterances? I meant to upload a bunch of new unlabeled utterances in order to annotate annotate or label for the first time, and then feed into the training data by “marking correct”.