LARS-research/AdaProp

Take the liberty to ask about the performance of other data sets(WebQSP)

Closed this issue · 1 comments

Has this model been run on the WebQuestionSP dataset? If so, how effective? If not, can predecessors roughly estimate the performance of the model from the differences between the NELL data set and the WebQSP data set?

Hi. We have not run the WebQuestionSP dataset yet. You are recommended to modify this code to fit your needs, and then you will get the empirical result.