INDArray的Java double损失了精度

Question

我不能将double转换为Nd4j.INDArray而不会丢失精度。

double[] weight = new double[]{-0.13404223866376802,-0.11294084872465669,0.11232944517596409,-0.01687720880184701,0.13288394029423414,-0.06916641552096867,0.1518882606786481};
INDArray array = Nd4j.create(weight, new int[]{1, 7});
System.out.println(array);

输出是

[-0.13,  -0.11,  0.11,  -0.02,  0.13,  -0.07,  0.15]

期望的输出应该是

[-0.13404223866376802,-0.11294084872465669,0.11232944517596409,-0.01687720880184701,0.13288394029423414,-0.06916641552096867,0.1518882606786481]

如何完全转换？

Answer 1

使用

int precision = 7;
String printed = new NDArrayStrings(precision).format(this);
System.out.println(printed);

无论你想要什么精度。 NDArrayStrings是org.nd4j.linalg.string包。

Answer 2

我有同样的问题，最新的0.9.x nd4j：

    new NDArrayStrings(15).format(Nd4j.create(new double[]{0.4,0.3}))
    [0.400000005960464,  0.300000011920929]

    Nd4j.create(new double[]{0.4}).getDouble(0)
    0.4000000059604645

编辑：现在我看到原因，Nd4j只初始化为“浮动”数字...所以只有7位数是正确的，最好使用“getFloat”

    Nd4j.create(new double[]{0.4 }).getFloat(0)
    0.4