Possible problem with reproducibility about reservoirpy HOT 1 CLOSED

Hello @jhklarcher, thank you for your interest in reservoirpy!

The difference you get is very small, in the order of magnitude of 10^-12. This sounds more related to the machine precision in representing floating numbers.

If it was a problem from the seed, chaning the seed would produce a change in the same order of magnitude.
If execute your code with different seeds, that's what I obtain:

0.008992516477587574
0.008992516477486478
0.008992516477550509
0.008992516477582382
0.008992516477508403
0.008992516477443787
0.008992516477566269
0.008992516477366569
0.008992516477584988
0.008992516477490897

0.010148570924057914
0.010148570923515286
0.010148570923515702
0.01014857092381421
0.010148570924057914
0.010148570923146085
0.010148570923329131
0.010148570923503379
0.010148570923438358
0.010148570923484517

0.010669491779120306
0.010669491778865479
0.010669491779081858
0.010669491778857215
0.010669491779142212
0.010669491778962882
0.010669491778869892
0.010669491779084567
0.010669491778886101
0.010669491778861293

If now I use the same seed as the last one, that's what I obtain:
0.010669491779120306
0.010669491778865479
0.010669491779081858
0.010669491778857215
0.010669491779142212
0.010669491778962882
0.010669491778869892
0.010669491779084567
0.010669491778886101
0.010669491778861293

Which is exactly the same output as before ... so the seed management is working! ;)

from reservoirpy.