Missing evidence selection model?

[sefinch]

The scripts/ README indicates that we should be able to perform evidence sentence selection on retrieved documents:

To run evidence sentence selection after document retrieval, use the get_evidence_scores function from the bert_evidence_selection.py.

It looks like the bert_evidence_selection code assumes a checkpoint is given to the evidence selection model, and if None, then it refers to the checkpoint like this:

if self.checkpoint is None:
    self.checkpoint = '/export/home/code/common/gitrepos/KernelGAT/checkpoint/retrieval_modelft_wowctxutt/model.best.pt'

Is there a link to the evidence selection model that I am just overlooking? Thanks