Errors choice
Posted: Tue 30 Mar 2021 12:11
Dear Didier,
during last error study I saw quite often at output screen the line ... [N][1]=+1.000 Error!!!
It is changed is some seconds with the typical report about current N-th run (see pics).
In parallel, the historgam for the losses is sometimes a bit strange, see the plot. I would expect it more Gaussian. At the example I have just a quadrupole displacement dx, dy.
Maybe the errors are now chosen not so uniformly?
Thank you and best regards,
Anna
during last error study I saw quite often at output screen the line ... [N][1]=+1.000 Error!!!
It is changed is some seconds with the typical report about current N-th run (see pics).
In parallel, the historgam for the losses is sometimes a bit strange, see the plot. I would expect it more Gaussian. At the example I have just a quadrupole displacement dx, dy.
Maybe the errors are now chosen not so uniformly?
Thank you and best regards,
Anna