Description
Describe the workflow you want to enable
Dear all,
when using HalvingGridSearch it is very likely, that (especially the first iteration) it contains 100's to 1000's of grid points. While monitoring in the verbose mode can be nice to get an idea about parameter importance, the only way to get an idea about the progress is to 'count rows'.
Describe your proposed solution
To get a better idea on the progress it would be nice to add the counter in the verbose = 3 output:
e.g. instead of
"[CV 1/3] END C=5, class_weight=None, degree=2, gamma=1, kernel=rbf;, score=(train=-0.050, test=-0.201) total time= 13.5s"
"[Iter 0; Fit Number 1234 of 4321; CV 1/3] END C=5, class_weight=None, degree=2, gamma=1, kernel=rbf;, score=(train=-0.050, test=-0.201) total time= 13.5s"
Both, Iter and total number of fits in the current iteration are provided at the beginning of each iteration and thus would require only minor changes.
Describe alternatives you've considered, if relevant
No response
Additional context
No response