Repo for Lab and Lectures for Deep Learning in Search from AFIRM 2020
The data used for the LTR tutorial is coming from Microsoft Learning to Rank Datasets. The data is in the following format: label qid features where the labels are a score 0-4(0 (irrelevant) to 4 (perfectly relevant)), qid is a number representing query and features is a 136-dimensional feature vector. To learn about what each feature represents check out the original dataset source.