Muting the remaining messages when show_progress=False #827
Conversation
|
Sorry for the delayed reply. Let us reuse the verbose_eval parameter |
|
Ok I see, I changed it! is it ok now? The travis tests are failing because I changed the name of the parameters, but the code should work correctly |
|
please do a rebase to merge your commits, and it will be merged.http://xgboost.readthedocs.org/en/latest/dev-guide/contribute.html#how-to-combine-multiple-commits-into-one |
|
done :D! |
Muting the remaining messages when show_progress=False
|
@ivallesp It seems that I merged the change too hastely and now the travis test break due to the change introduced by this PR, https://travis-ci.org/dmlc/xgboost/builds/110415766 Please make a PR to fix this issue, so the other changes can be savely merged in. |
|
@tqchen I just did it! Sorry but I made something wrong during the squash and the changes in the training file were not applied. I think it is solved now. Let's wait until tests done. Thank you for reporting it. |
|
@tqchen yep! Now the tests passed and it is working correctly. Sorry for the inconvenience. |
I was using the library and I found that when doing a cross validation, even when I chose the parameter show_progress to False, it showed some messages:
"Will train until cv error hasn't decreased in X rounds", in case of having early_stopping
"Stopping. Best iteration:\n[X] cv-mean:[...]\n"
I wrote some conditions for muting it when show_progress=Fase because if you run a bunch of models with a grid of parameters, the high number of messages can bother :D.