CLMBR Fails to Create Batches for MIMIC-IV Data

[Ghebes]

Good evening,

To compare models, I wanted to test MIMIC-IV and EHRSHOT on CLMBR and view the performance. Whenever I run through the scripts with the EHRSHOT data, every script works; however, whenever I run the clmbr_create_batches script with MIMIC, the program aborts without any feedback. I tested twice with EHRSHOT to guarantee that none of the code was being changed, and likewise altered the batch_size to another number that was a power of 2. At first, I thought that it might be due to some of the codes not being converted to accepted codes in CLMBR dictionary, but after removing these rows from the dataset and rerunning the whole script, the same aborted result happens in the clmbr_create_batches. When looking to determine the location of the error, I found that it was the femr.extension.dataloader.create_batches method, which is not available on any of the versions of this GitHub repo. I need to be able to review the create_batches method in order to determine what might be causing this error. Is there a way that I can do this?