Question about data split?

[tdzdog]

Can the code for splitting the data be provided? I am aiming to split our custom dataset. But there is still some confusion after reading the paper. To be specific, you split the whole data into four tasks. But are there categories overlapping among the tasks? For example, are instances of task-2 categories in task 1 during training? Or are there instances of task-1 categories in following tasks 2, 3, and 4?