I'm using GridSearchCV to do some analysis for a model (same problem happens with learning_curve btw) and have verbose parameter set to 2. The only output I get in the notebook is some minor progress logs at seemingly arbitrary points and it just prints 'Done' when the kernel gets idle.
If I use the default value for n_jobs (1) then the output is properly shown in the notebook but with parallel processing this happens:
The detailed log is printed in the console while the kernel is running.
Is there any way to have the detailed progress log be printed in the notebook?
If there is a nicer way of getting real-time progress logs from sklearn (maybe something with tqdm), that would work too.