Sd.evaluate failure
adonnini opened this issue · 0 comments
adonnini commented
When I run the code reported below, execution fails with
Index can't be < 0 and >= 1984
What am I missing / doing wrong? Thanks
CODE
String outputVariable = outReduced.name();
evaluation = new Evaluation();
testData.reset();
sd.evaluate(testData, outputVariable, evaluation);
outReduced
has shape [64, 2, 31]
testData
has 109 elements.
ERROR LOG
Exception in thread "main" java.lang.IllegalStateException: Index can't be < 0 and >= 1984 at org.nd4j.linalg.factory.Nd4j.pullRows(Nd4j.java:4749) at org.nd4j.linalg.factory.Nd4j.pullRows(Nd4j.java:4722) at org.nd4j.evaluation.EvaluationUtils.extractNonMaskedTimeSteps(EvaluationUtils.java:208) at org.nd4j.evaluation.BaseEvaluation.reshapeAndExtractNotMasked(BaseEvaluation.java:186) at org.nd4j.evaluation.classification.Evaluation.eval(Evaluation.java:361) at org.nd4j.evaluation.BaseEvaluation.eval(BaseEvaluation.java:267) at org.nd4j.autodiff.samediff.SameDiff.evaluateHelper(SameDiff.java:2402) at org.nd4j.autodiff.samediff.SameDiff.evaluate(SameDiff.java:2296) at org.nd4j.autodiff.samediff.config.EvaluationConfig.exec(EvaluationConfig.java:197) at org.nd4j.autodiff.samediff.SameDiff.evaluate(SameDiff.java:2206) at org.deeplearning4j.examples.trajectorypredictiontransformer.LocationNextNeuralNetworkV7_04.fitAndEvaluateTestDataset(LocationNextNeuralNetworkV7_04.java:1083) at org.deeplearning4j.examples.trajectorypredictiontransformer.LocationNextNeuralNetworkV7_04.sameDiff3(LocationNextNeuralNetworkV7_04.java:685) at org.deeplearning4j.examples.trajectorypredictiontransformer.LocationNextNeuralNetworkV7_04.main(LocationNextNeuralNetworkV7_04.java:207)