Is the performance described in the paper based on the DKT1 model?

[AkihikoWatanabe]

Hello. I have learned various insights from your work. Thank you for your great work. I have a question related to the DKT performance reported in your paper, "When is Deep Learning the Best Approach to Knowledge Tracing?".

Looking at the DKT implementation in this repository, two types of DKT exist, DKT1 and DKT2. So my question is which of the DKT implementations, DKT1 or DKT2, is responsible for the DKT performance described in the paper (i.e. the performance described in Table3 ~ Table11 and Table14)?

Thank you very much.