Why does __get_item__ use the same index for both distributions? Shouldn't you pick 2 random indexes. Otherwise you have a constant pairing between the two distributions.
Thank you for your question! Yes, the pseudocode involve creating a pairing of the data and sticking with it. If you use reflow to straighten the mapping, the reflow operations build new pairings. It depends on papers and implementations whether _get_item_ uses the same index or no.
Why does __get_item__ use the same index for both distributions? Shouldn't you pick 2 random indexes. Otherwise you have a constant pairing between the two distributions.
Thank you for your question! Yes, the pseudocode involve creating a pairing of the data and sticking with it. If you use reflow to straighten the mapping, the reflow operations build new pairings. It depends on papers and implementations whether _get_item_ uses the same index or no.