swan343

hi @tbepler thanks for the response! From my limited testing I do see the results are repeatable with multiprocessing, if the seed argument is given (np.random.RandomState should always return the...

@tbepler thanks for the reply, by workers are you referring to `--num-workers` or something else? I can generate the exact same output with --num-workers range from 0 to 9 from...

yes that's exactly the intent. The variability isn't a lot per se - I do particle masking and can see about 2%-5% of difference depending on the detection technique and...