This is to ensure correct indices, right? And pm.tt_rng somehow generates different sample at each iteration, but not each time it is used in code?
Also, do you mean data[ridx][maskshared[ridx]] in the last line?
This is to ensure correct indices, right? And pm.tt_rng somehow generates different sample at each iteration, but not each time it is used in code?
Also, do you mean data[ridx][maskshared[ridx]] in the last line?