Joey Gao
Joey Gao
Really needed, as 'transform` is often been used in `groupby` operations!
same features needed, any update?
same errors, the problem lies in : ```python def _prediction_feature_weights(booster, dmatrix, n_targets, feature_names, xgb_feature_names): """ For each target, return score and numpy array with feature weights on this prediction, following...
Now the question is: Is `dmatrix` only contains one sample ?
这样来说也是，early stopping是不必要，不过像ligthgbm/catboost的categorical_feature特征应该是要用的吧，cat/lgb对cat特征的处理有的时候是优于一般的label encoder处理的
`1.5.1`版本的应该可以吧，最近的没有测试了
Did you import the module by `from epolot import eplot` or `import eplot` successfully?
`groupby` and `apply` is in an urgent need. Any progress?
such as every group's `head` n rows, or every group `size`/`nlargest`..., all this can get by `df.groupby('_id').apply(lambd x:x.head(N))` or `df.groupby('_id').apply(lambd x:x.shape[0])`
> ly 请问这里`NoMask`有什么作用啊，去掉会不会有影响？