You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'd like to follow some parameters during training and hence every statinterval iterations plot their evolutions up to that point. Here is how I decided to do that:
if math.fmod(t, statinterval) == 0 then
f1= gnuplot.figure()
gnuplot.plot({'Param', torch.Tensor(paramlist), '-'})
end
To avoid having thousands of plots, I tried to add gnuplot.close() or gnuplot.close(f1) before creating a new plot but it does not have any consequences. Is this a bug or am I misusing gnuplot.close? Thanks!
Update
Picking a plot id makes the plot be interactively updated (ie f = gnuplot.figure(1)). Still, the close function does not seem to work although I don't really need it anymore.
The text was updated successfully, but these errors were encountered:
Hi,
I'd like to follow some parameters during training and hence every
statinterval
iterations plot their evolutions up to that point. Here is how I decided to do that:To avoid having thousands of plots, I tried to add
gnuplot.close()
orgnuplot.close(f1)
before creating a new plot but it does not have any consequences. Is this a bug or am I misusing gnuplot.close? Thanks!Update
Picking a plot id makes the plot be interactively updated (ie
f = gnuplot.figure(1)
). Still, theclose
function does not seem to work although I don't really need it anymore.The text was updated successfully, but these errors were encountered: