使用MultilayerPerceptronClassifier和ML库的数据类型错误

时间:2019-06-23 15:01:49

标签: apache-spark pyspark

我想为Shader "selectedFace" { SubShader { Pass { CGPROGRAM #pragma vertex vert #pragma fragment frag //for maximum portability using the UNITY_VPOS_TYPE **#pragma target 3.0** float4 vert(float4 vertexPosition:POSITION):POSITION { return mul(UNITY_MATRIX_MVP,vertexPosition); } //Mapping SV_Target(D3D10) to COLOR(Direct3D 9) **float4 frag(UNITY_VPOS_TYPE screenPos : VPOS):SV_Target** { //screenPos.x - x position on Screen //screenPos.y - y position on Screen return float4(0.75,1.0,0.0,1.0); } ENDCG } } } 库中的MultilayerPerceptronClassifier添加一些数据,以进行整体堆叠:

PySpark ML

当我运行上面的代码时,出现此错误:

  

AttributeError:'numpy.ndarray'对象没有属性'_jdf'

我试图将mlp = MultilayerPerceptronClassifier(maxIter=10, layers=[5,3], blockSize=128) features = features_vector model_stacking = mlp.fit(features) predictions_foo = model_cv.transform(validation) predictions = evaluator.evaluate(predictions_foo) 的类型从np.array更改为Pandas Dataframe,但是我遇到了同样的错误。

此分类器接受哪种类型的数据?

0 个答案:

没有答案