编译deeplearning4j出错

错误如下：
[ERROR] /home/luolianhai/deeplearning4j/deeplearning4j-scaleout/spark/dl4j-spark-ml/src/main/scala/org/deeplearning4j/spark/ml/nn/TrainingStrategy.scala:79: overloaded method value zeros with alternatives:
(x$1: [Int])org.nd4j.linalg.api.ndarray.INDArray
(x$1: Int)org.nd4j.linalg.api.ndarray.INDArray
cannot be applied to (Array[Int])
[ERROR] val accumulatedParams = sc.accumulator(Nd4j.zeros(networkParams.shape()))(INDArrayAccumulatorParam)
[ERROR] ^
[ERROR] /home/luolianhai/deeplearning4j/deeplearning4j-scaleout/spark/dl4j-spark-ml/src/main/scala/org/deeplearning4j/spark/ml/nn/TrainingStrategy.scala:103: overloaded method value zeros with alternatives:
(x$1: [Int])org.nd4j.linalg.api.ndarray.INDArray
(x$1: Int)org.nd4j.linalg.api.ndarray.INDArray
cannot be applied to (Array[Int])
[ERROR] def zero(initialValue: INDArray): INDArray = Nd4j.zeros(initialValue.shape())
[ERROR] ^
[ERROR] two errors found
[INFO] ---------------------------------

TrainingStrategy.scala的第79和103行如下：
val accumulatedParams = sc.accumulator(Nd4j.zeros(networkParams.shape()))(INDArrayAccumulatorParam)
def zero(initialValue: INDArray): INDArray = Nd4j.zeros(initialValue.shape())

#include
int main()
{
int a,b;
double c;
scanf("%d%d",&a,&b);
c=100.0*b/a;
printf("%.3lf%%",c);
return 0;
}

spark 貌似不支持直接支持 深度学习吧， 你可以通过 deeplearning4j与spark整合来支持。