无法创建DaskDMatrix

Question

我正在尝试按照文档使用 dask 训练 xgboost 模型。我的问题是我在必须创建 DaskDMatrix 的步骤中被阻止。无论我如何尝试，我都会收到一条错误消息，表明该方法不存在。我试过：

import dask
import dask_xgboost as dxgb

client = dask.distributed.Client()

params = {'objective':'binary:logistic', 
        'booster': 'dart', 
          'n_estimators': 800,
          'max_depth': 4,
          'learning_rate': 0.02, 
          'random_state': 42}
}

dtrain = dxgb.DaskDMatrix(client, X_train, y_train)
dval = dxgb.DaskDMatrix(client, X_val, y_val)

eval_set = [(dtrain, 'train'), (dval, 'validation')]

model = dxgb.train(client, params, dtrain, num_boost_round=10, evals=eval_set, eval_metric=['logloss', 'aucpr'], verbose=True)

==>“属性错误：模块‘dask_xgboost’没有属性‘DaskDMatrix’”

然后我尝试了：

import xgboost as xgb

# Create a Dask DMatrix from a Dask DataFrame
dtrain = xgb.dask.DaskDMatrix(client, X_train, y_train)

==> 属性错误：模块“xgboost”没有属性“dask”

还有：

import dask_ml.xgboost as dxgb

# Create a Dask DMatrix from a Dask DataFrame
dtrain = dxgb.DaskDMatrix(client, X_train, y_train)

==> 属性错误：模块“dask_ml.xgboost”没有属性“DaskDMatrix”

在哪里可以找到正确的代码？

Answer 1

在我的笔记本电脑上使用新安装的 conda xgboost 和 dask-xgboost 运行代码，并遵循 文档：

import xgboost as xgb
import dask.array as da
import dask.distributed

client = dask.distributed.Client()

params = {'objective':'binary:logistic', 
        'booster': 'dart', 
          'n_estimators': 800,
          'max_depth': 4,
          'learning_rate': 0.02, 
          'random_state': 42}

X_train = da.random.random((10,100000))
y_train = da.random.random((1,100000))
dtrain = xgb.dask.DaskDMatrix(client, X_train, y_train)