1. Test Modules
  2. Differential Validation
    1. Feedback Validation
    2. Learning Validation
    3. Total Accuracy
    4. Frozen and Alive Status
  3. Results

Subreport: Logs for com.simiacryptus.ref.lang.ReferenceCountingBase

Test Modules

Using Seed 8154491105761915904

Differential Validation

SingleDerivativeTester.java:101 executed in 0.00 seconds (0.000 gc):

        log.info(RefString.format("Inputs: %s", prettyPrint(inputPrototype)));
        log.info(RefString.format("Inputs Statistics: %s", printStats(inputPrototype)));
        log.info(RefString.format("Output: %s", outputPrototype.prettyPrint()));
        assert outputPrototype != null;
        log.info(RefString.format("Outputs Statistics: %s", outputPrototype.getScalarStatistics()));
      },
      outputPrototype.addRef(),
      RefUtil.addRef(inputPrototype)));
Logging
Inputs: [
[ [ 0.08 ], [ -0.128 ] ],
[ [ 0.7 ], [ 0.496 ] ]
],
[
[ [ -0.608 ], [ 0.048 ] ],
[ [ 1.764 ], [ 1.524 ] ]
]
Inputs Statistics: {meanExponent=-0.6122800817139336, negative=1, min=-0.128, max=0.7, mean=0.287, count=4, sum=1.148, positive=3, stdDev=0.3276141022605712, zeros=0},
{meanExponent=-0.27634290888807383, negative=1, min=-0.608, max=1.764, mean=0.682, count=4, sum=2.728, positive=3, stdDev=0.993194844932252, zeros=0}
Output: [
[ [ -0.527999997138977 ], [ -0.08000000566244125 ] ],
[ [ 2.4639999866485596 ], [ 2.0199999809265137 ] ]
]
Outputs Statistics: {meanExponent=-0.16932099890073332, negative=2, min=-0.527999997138977, max=2.4639999866485596, mean=0.9689999911934137, count=4, sum=3.875999964773655, positive=2, stdDev=1.2923849971371364, zeros=0}

Feedback Validation

We validate the agreement between the implemented derivative of the inputs apply finite difference estimations:

SingleDerivativeTester.java:117 executed in 0.08 seconds (0.000 gc):

        return testFeedback(
            statistics,
            component.addRef(),
            RefUtil.addRef(inputPrototype),
            outputPrototype.addRef());
      },
      outputPrototype.addRef(),
      RefUtil.addRef(inputPrototype),
      component.addRef()));
Logging
Feedback for input 0
Inputs Values: [
[ [ 0.08 ], [ -0.128 ] ],
[ [ 0.7 ], [ 0.496 ] ]
]
Value Statistics: {meanExponent=-0.6122800817139336, negative=1, min=-0.128, max=0.7, mean=0.287, count=4, sum=1.148, positive=3, stdDev=0.3276141022605712, zeros=0}
Implemented Feedback: [ [ 1.0, 0.0, 0.0, 0.0 ], [ 0.0, 1.0, 0.0, 0.0 ], [ 0.0, 0.0, 1.0, 0.0 ], [ 0.0, 0.0, 0.0, 1.0 ] ]
Implemented Statistics: {meanExponent=0.0, negative=0, min=0.0, max=1.0, mean=0.25, count=16, sum=4.0, positive=4, stdDev=0.4330127018922193, zeros=12}
Measured Feedback: [ [ 1.000046730041504, 0.0, 0.0, 0.0 ], [ 0.0, 1.0001659393310547, 0.0, 0.0 ], [ 0.0, 0.0, 1.0000020265579224, 0.0 ], [ 0.0, 0.0, 0.0, 1.0001659393310547 ] ]
Measured Statistics: {meanExponent=4.132384032299968E-5, negative=0, min=0.0, max=1.0001659393310547, mean=0.250023789703846, count=16, sum=4.000380635261536, positive=4, stdDev=0.4330539083861724, zeros=12}
Feedback Error: [ [ 4.673004150390625E-5, 0.0, 0.0, 0.0 ], [ 0.0, 1.659393310546875E-4, 0.0, 0.0 ], [ 0.0, 0.0, 2.0265579223632812E-6, 0.0 ], [ 0.0, 0.0, 0.0, 1.659393310546875E-4 ] ]
Error Statistics: {meanExponent=-4.395936535534613, negative=0, min=0.0, max=1.659393310546875E-4, mean=2.3789703845977783E-5, count=16, sum=3.8063526153564453E-4, positive=4, stdDev=5.488870942214264E-5, zeros=12}
Feedback for input 1
Inputs Values: [
[ [ -0.608 ], [ 0.048 ] ],
[ [ 1.764 ], [ 1.524 ] ]
]
Value Statistics: {meanExponent=-0.27634290888807383, negative=1, min=-0.608, max=1.764, mean=0.682, count=4, sum=2.728, positive=3, stdDev=0.993194844932252, zeros=0}
Implemented Feedback: [ [ 1.0, 0.0, 0.0, 0.0 ], [ 0.0, 1.0, 0.0, 0.0 ], [ 0.0, 0.0, 1.0, 0.0 ], [ 0.0, 0.0, 0.0, 1.0 ] ]
Implemented Statistics: {meanExponent=0.0, negative=0, min=0.0, max=1.0, mean=0.25, count=16, sum=4.0, positive=4, stdDev=0.4330127018922193, zeros=12}
Measured Feedback: [ [ 0.9999871253967285, 0.0, 0.0, 0.0 ], [ 0.0, 0.9999275207519531, 0.0, 0.0 ], [ 0.0, 0.0, 0.9999945759773254, 0.0 ], [ 0.0, 0.0, 0.0, 0.9999275207519531 ] ]
Measured Statistics: {meanExponent=-1.7725997794855715E-5, negative=0, min=0.0, max=0.9999945759773254, mean=0.2499897964298725, count=16, sum=3.99983674287796, positive=4, stdDev=0.4329950290818009, zeros=12}
Feedback Error: [ [ -1.2874603271484375E-5, 0.0, 0.0, 0.0 ], [ 0.0, -7.2479248046875E-5, 0.0, 0.0 ], [ 0.0, 0.0, -5.424022674560547E-6, 0.0 ], [ 0.0, 0.0, 0.0, -7.2479248046875E-5 ] ]
Error Statistics: {meanExponent=-4.608879322599185, negative=4, min=-7.2479248046875E-5, max=0.0, mean=-1.0203570127487183E-5, count=16, sum=-1.6325712203979492E-4, positive=0, stdDev=2.3764275621071893E-5, zeros=12}

Returns

    {
      "absoluteTol" : {
        "count" : 32,
        "sum" : 5.438923835754395E-4,
        "min" : 0.0,
        "max" : 1.659393310546875E-4,
        "sumOfSquare" : 6.796118512397697E-8,
        "standardDeviation" : 4.283574869505047E-5,
        "average" : 1.6996636986732483E-5
      },
      "relativeTol" : {
        "count" : 8,
        "sum" : 2.719345535710119E-4,
        "min" : 1.0132779344484278E-6,
        "max" : 8.296278213305896E-5,
        "sumOfSquare" : 1.6988177371073696E-8,
        "standardDeviation" : 3.1113958268035633E-5,
        "average" : 3.3991819196376485E-5
      }
    }

Learning Validation

We validate the agreement between the implemented derivative of the internal weights apply finite difference estimations:

SingleDerivativeTester.java:133 executed in 0.00 seconds (0.000 gc):

        return testLearning(
            statistics,
            component.addRef(),
            RefUtil.addRef(inputPrototype),
            outputPrototype.addRef());
      },
      outputPrototype.addRef(),
      RefUtil.addRef(inputPrototype),
      component.addRef()));

Returns

    {
      "absoluteTol" : {
        "count" : 32,
        "sum" : 5.438923835754395E-4,
        "min" : 0.0,
        "max" : 1.659393310546875E-4,
        "sumOfSquare" : 6.796118512397697E-8,
        "standardDeviation" : 4.283574869505047E-5,
        "average" : 1.6996636986732483E-5
      },
      "relativeTol" : {
        "count" : 8,
        "sum" : 2.719345535710119E-4,
        "min" : 1.0132779344484278E-6,
        "max" : 8.296278213305896E-5,
        "sumOfSquare" : 1.6988177371073696E-8,
        "standardDeviation" : 3.1113958268035633E-5,
        "average" : 3.3991819196376485E-5
      }
    }

Total Accuracy

The overall agreement accuracy between the implemented derivative and the finite difference estimations:

SingleDerivativeTester.java:148 executed in 0.00 seconds (0.000 gc):

    //log.info(String.format("Component: %s\nInputs: %s\noutput=%s", component, Arrays.toStream(inputPrototype), outputPrototype));
    log.info(RefString.format("Finite-Difference Derivative Accuracy:"));
    log.info(RefString.format("absoluteTol: %s", statistics.absoluteTol));
    log.info(RefString.format("relativeTol: %s", statistics.relativeTol));
Logging
Finite-Difference Derivative Accuracy:
absoluteTol: 1.6997e-05 +- 4.2836e-05 [0.0000e+00 - 1.6594e-04] (32#)
relativeTol: 3.3992e-05 +- 3.1114e-05 [1.0133e-06 - 8.2963e-05] (8#)

Frozen and Alive Status

SingleDerivativeTester.java:156 executed in 0.01 seconds (0.000 gc):

    testFrozen(component.addRef(), RefUtil.addRef(inputPrototype));
    testUnFrozen(component.addRef(), RefUtil.addRef(inputPrototype));

LayerTests.java:425 executed in 0.00 seconds (0.000 gc):

    throwException(exceptions.addRef());

Results

classdetailsresult
com.simiacryptus.mindseye.test.unit.SingleDerivativeTesterToleranceStatistics{absoluteTol=1.6997e-05 +- 4.2836e-05 [0.0000e+00 - 1.6594e-04] (32#), relativeTol=3.3992e-05 +- 3.1114e-05 [1.0133e-06 - 8.2963e-05] (8#)}OK
  {
    "result": "OK",
    "performance": {
      "execution_time": "0.249",
      "gc_time": "0.122"
    },
    "created_on": 1586741888803,
    "file_name": "derivativeTest",
    "report": {
      "simpleName": "Float_Add",
      "canonicalName": "com.simiacryptus.mindseye.layers.cudnn.SumInputsLayerTest.Float_Add",
      "link": "https://github.com/SimiaCryptus/mindseye-cudnn/tree/59d5b3318556370acb2d83ee6ec123ce0fc6974f/src/test/java/com/simiacryptus/mindseye/layers/cudnn/SumInputsLayerTest.java",
      "javaDoc": ""
    },
    "archive": "s3://code.simiacrypt.us/tests/com/simiacryptus/mindseye/layers/cudnn/SumInputsLayer/Float_Add/derivativeTest/202004133808",
    "id": "3533443a-1c51-4ec1-a0ba-3a30d28e0eb8",
    "report_type": "Components",
    "display_name": "Derivative Validation",
    "target": {
      "simpleName": "SumInputsLayer",
      "canonicalName": "com.simiacryptus.mindseye.layers.cudnn.SumInputsLayer",
      "link": "https://github.com/SimiaCryptus/mindseye-cudnn/tree/59d5b3318556370acb2d83ee6ec123ce0fc6974f/src/main/java/com/simiacryptus/mindseye/layers/cudnn/SumInputsLayer.java",
      "javaDoc": ""
    }
  }