Lightgbm: [LightGBM] [Info] No further splits with positive gain, best gain: -inf

[zyoohv]

Environment info

Operating System: Ubuntu16.04
CPU:Intel
Python version: Python 2.7

Error Message:

Could you tell me what's the mean of this message?

[LightGBM] [Info] No further splits with positive gain, best gain: -inf
[LightGBM] [Info] Trained a tree with leaves=58 and max_depth=14

Param

    "param": {  
    "task":"train",  
    "boosting":"gbdt",  
    "application":"binary",  
    "num_leaves":63,  
    "min_data_in_leaf":1000,  
    "learning_rate":0.1,  
    "feature_fraction":1.0,  
    "bagging_freq":50,  
    "bagging_fraction":0.8  
    }

[guolinke]

@zyoohv
it means the learning of tree in current iteration should be stop, due to cannot split any more.
I think this is caused by "min_data_in_leaf":1000, you can set it to a smaller value.

[zyoohv]

@guolinke
Thank you very much. I changed it according your recommend. And now it works! :)

[guolinke]

@zyoohv
This is not a bug, it is a feature.
The output message is to warn user that your parameters may be wrong, or your dataset is not easy to learn.

[AdamsMi]

And how to disable this 'feature', if it's possible :)?

[guolinke]

you can shut up the warning by set verbose

for sklearn interface, you can set verbose=-1 when defining the model (not in fit).
for lgb.train interface, you can set verbose=-1 in param dict.

[xuwiliam]

I also meet this problem, but can not deal with it using the solution referenced in issue640

[guolinke]

@xuwiliam it should work if you on the latest master branch.
If it still doesn't work, you can provide a re-produce code with random generated input data.

[goldentom42]

Hi there, the problem still occurs even with the latest master.

Using lgb.train and verbose set to -1 you don't get the warning if and only if you don't provide an argument for valid_sets. If valid_sets is provided you can set verbose or verbosity to whatever value the warning will be displayed.

Hope this helps in finding the issue.

[wxkpythonwork]

set lgb.LGBMRegressor(verbose=-1), it works!