yseop / yseoplab

PUBLIC PROJECTS

yseoplab's Issues

Task2 baseline - repeated instances after preprocessing

Hi,

first, thank you for the updates. The baselines and especially the new scoring functions are very useful.

When I started debugging the task2_baseline.py I think I found a bug. There is a huge amount of repeated instances in the data variable, which is the input of the train and test datasets. In the following picture, there are 9 occurrences of the instance which start with: "It found that total U.S. ...".

I found the problem should be in somewhere in the make_causal_input function because the input lodict_ contains 641 instances, but the length of the output (hometags) is 205761.