This is the supplementary material of ExerCAKT. The appendix has been included in the PDF and uploaded together with the text. 
Dataset acquisition reference pyKT. The supplementary material mainly includes the code involved in this article. The code path corresponding to our model is "openreviewcode  pykt  models  akt. py" (due to time constraints, we temporarily borrowed the AKT pipeline, which will be adjusted in the official Github code). 
The configuration of ablation experiments has been included, and performance comparison and super parameter experiments are completed through the sweep of the wandb library. 
It only takes three steps to run the code：(1) Install wandb and other required libraries (compared with pyKT, we have not introduced other libraries); 
(2) Configure the settings of sweep in wandb (in the file sweepconfig. txt); 
(3) Use sweep_ExerCAKT_train.py for training and visualization as required by wandb. 
You can also implement training according to the pyKT native method.
Thank you！