Commenting use_label_encoder - xgboost #1122
Conversation
sonichi
requested review from
markharley,
thinkall,
qingyun-wu,
ZviBaratz,
coffepowered and
EgorKraevTransferwise
|
@qingyun-wu Check benchmark option. |
|
@minghao51, thank you for the contribution. We have some installation option that requires xgboost version earlier than 1.7.0. Can you please first check the version of xgboost and only comment the line only when the version number is larger than or equal to 1.7.0? This ensure the code still works fine even when an earlier version of xgboost is used. Thank you! |
|
Sure |
|
Updated the PR |
flaml/automl/model.py
Outdated
| params["use_label_encoder"] = params.get("use_label_encoder", False) | ||
| # params["booster"] = params.get("booster", "gbtree") | ||
| # use_label_encoder is deprecated in 1.7.0 | ||
| from xgboost import __version__ as xgboost_version |
There was a problem hiding this comment.
This code block should not be under the max_depth == 0 condition.
There was a problem hiding this comment.
should the import xgboost version be under the function then? or at the very start?
flaml/automl/model.py
Outdated
| if xgboost_version < "1.7.0": | ||
| params["use_label_encoder"] = params.get("use_label_encoder", False) |
There was a problem hiding this comment.
Sorry about the confusion. I mean all these several lines of code.
There was a problem hiding this comment.
They need to be outside of the condition if max_depth == 0: according to the original logic.
github-merge-queue
bot
removed this pull request from the merge queue due to failed status checks
Why are these changes needed?
use_label_encoderis deprecated since 1.7.0. (xgboost), yet it is still currently provided as part of the hyperparameter for xgboostRelated issue number
#1120
Checks