OLR2020 dataset preparation

[alumae]

The recipe asv-subtools/recipe/ap-olr2020-baseline/run.sh uses datasets like data/task1_test, data/task2_enroll, etc. While it's mostly easy to understand what data they correspond to by reading the paper "AP20-OLR Challenge: Three Tasks and Their
Baselines" (https://arxiv.org/pdf/2006.03473.pdf), some things are confusing. For example, what exactly data/train contains, and data/task1_enroll. perhaps you can publish the script that was used for creating the data directories?