Open perprit opened 3 years ago
@perprit , Hi, Thanks very much for your question. The SE score was calculated with the model trained by "Step 1. SE Training". There is no direct table to reflect the score because it is calculated dynamically. So, you should feed in some text and the model can calculate the score. Unfortunately, we get this image by hard code into the process and there is no quick script to do this. We are sorry for not considering this.
Hi, thanks for the great work!
I read the paper and ran the scripts, and I wonder if it is possible to get the visualization of the supportiveness estimator score for each source KB-target text pair.
Can I get the image by running a script? if that's not possible, how can I get the trained SE score?
Thanks for the reply in advance.