AutoGradSpec.scala

package scorch

import scorch.autograd._
import botkop.numsca.Tensor
import botkop.{numsca => ns}
import org.scalactic.{Equality, TolerantNumerics}
import org.scalatest.{FlatSpec, Matchers}

class AutoGradSpec extends FlatSpec with Matchers {
  
  "Autograd" should "derive mse" in {
    val nOut = 4
    val minibatch = 3

    val input  = Variable(ns.randn(minibatch, nOut))
    val label = Variable(ns.randn(minibatch, nOut))

    val diff = input - label
    val sqDiff = diff * diff
    val msePerEx = mean(sqDiff)
    val avgMSE = mean(msePerEx)

    avgMSE.shape shouldBe List(1, 1)

    avgMSE.backward()

    input.grad.get.shape shouldBe input.shape
  }

}

SameDiffTests.java

    @Test
    public void testMseBackwards() {

        SameDiff sd = SameDiff.create();

        int nOut = 4;
        int minibatch = 3;
        SDVariable input = sd.var("in", new int[]{-1,nOut});
        SDVariable label = sd.var("label", new int[]{-1, nOut});

        SDVariable diff = input.sub(label);
        SDVariable sqDiff = diff.mul(diff);
        SDVariable msePerEx = sd.mean("msePerEx", sqDiff, 1);
        SDVariable avgMSE = sd.mean("loss", msePerEx, 0);

        INDArray inputArr = Nd4j.rand(minibatch, nOut);
        INDArray labelArr = Nd4j.rand(minibatch, nOut);

        sd.associateArrayWithVariable(inputArr, input);
        sd.associateArrayWithVariable(labelArr, label);

        INDArray result = sd.execAndEndResult();
        assertEquals(1, result.length());

        Pair<Map<SDVariable, DifferentialFunction>, List<DifferentialFunction>> p = sd.execBackwards();
    }