Reputation: 1511
I was trying setup a single label labeling task in AWS Groundtruth through the console. My goal is to match some users in social media and for each user I have several possible candidates out of which one should be selected (label). My CSV looks like this:
firtname | lastname | candidates
Romeo Montague x
Juliet Capulet x
Instead of "x", I would like to have something like this
candidate_1
description_1
link_1
candidate_2
description_2
link_2
candidate_3
description_3
link_3
The human worked should then select whereas the correct label is candidate_1, candidate_2 or candidate_3 or none of the above.
I am aware Sagemaker ground truth does not accept new lines characters and that it renders it in HTML so I tried to input the following:
candidate_1 <br/> description <br/> link <br/><br/> candidate_2 <br/> description <br/> link <br/><br/> candidate_3 <br/> description <br/> link <br/><br/>
unfortunately, when I take a look at the console, the input on the left does not render correctly:
The line breaks within the div tag seem to be simply ignored by the UI.
I found this post which contains the answer but I am struggling to adapt to my concrete use case.
How can I change my csv so that the multiple paragraphs get rendered corrected?
Upvotes: 2
Views: 414