Index all 15 training sources: the nine own repos, the three DPR query sets, and the three BEIR training splits

index: 10 datasets

index: 9 datasets

index: 8 datasets

initial commit