ROCMultiClass error while running textCNN

DL4J
1

I replaced defaut Evaluation with ROCMultiClass in textCNN example.
EvaluativeListener eva = new EvaluativeListener(testIter, 1, InvocationType.EPOCH_END, new ROCMultiClass());
net.setListeners(new ScoreIterationListener(100), eva);
net.fit(trainIter, nEpochs);

Error occured when evaluating：
Exception in thread “main” java.lang.RuntimeException: cudaMalloc failed
at org.nd4j.jita.allocator.pointers.cuda.cudaStream_t.synchronize(cudaStream_t.java:41)
at org.nd4j.jita.handler.impl.CudaZeroHandler.alloc(CudaZeroHandler.java:307)
at org.nd4j.jita.allocator.impl.AtomicAllocator.allocateMemory(AtomicAllocator.java:526)
at org.nd4j.jita.allocator.impl.AtomicAllocator.allocateMemory(AtomicAllocator.java:438)
at org.nd4j.linalg.jcublas.buffer.BaseCudaDataBuffer.initPointers(BaseCudaDataBuffer.java:297)
at org.nd4j.linalg.jcublas.buffer.BaseCudaDataBuffer.(BaseCudaDataBuffer.java:316)
at org.nd4j.linalg.jcublas.buffer.CudaDoubleDataBuffer.(CudaDoubleDataBuffer.java:62)
at org.nd4j.linalg.jcublas.buffer.factory.CudaDataBufferFactory.create(CudaDataBufferFactory.java:378)
at org.nd4j.linalg.factory.Nd4j.createBuffer(Nd4j.java:1430)
at org.nd4j.linalg.jcublas.JCublasNDArrayFactory.create(JCublasNDArrayFactory.java:1543)
at org.nd4j.linalg.jcublas.JCublasNDArrayFactory.create(JCublasNDArrayFactory.java:1538)
at org.nd4j.linalg.factory.Nd4j.create(Nd4j.java:4298)
at org.nd4j.evaluation.classification.ROC.eval(ROC.java:626)
at org.nd4j.evaluation.BaseEvaluation.eval(BaseEvaluation.java:298)
at org.nd4j.evaluation.classification.ROCMultiClass.eval(ROCMultiClass.java:236)
at org.nd4j.evaluation.BaseEvaluation.eval(BaseEvaluation.java:308)
at org.deeplearning4j.nn.graph.ComputationGraph.doEvaluationHelper(ComputationGraph.java:4183)
at org.deeplearning4j.nn.graph.ComputationGraph.doEvaluationHelper(ComputationGraph.java:4123)
at org.deeplearning4j.nn.graph.ComputationGraph.doEvaluation(ComputationGraph.java:4081)
at org.deeplearning4j.nn.graph.ComputationGraph.doEvaluation(ComputationGraph.java:4068)
at org.deeplearning4j.optimize.listeners.EvaluativeListener.invokeListener(EvaluativeListener.java:243)
at org.deeplearning4j.optimize.listeners.EvaluativeListener.onEpochEnd(EvaluativeListener.java:213)
at org.deeplearning4j.nn.graph.ComputationGraph.fit(ComputationGraph.java:1090)
at org.deeplearning4j.nn.graph.ComputationGraph.fit(ComputationGraph.java:1018)
at org.deeplearning4j.nn.graph.ComputationGraph.fit(ComputationGraph.java:1006)
at org.deeplearning4j.examples.convolution.sentenceclassification.TextCNN4MultiIntentCommand.main(TextCNN4MultiIntentCommand.java:170)

by the way，https://community.konduit.ai is not easy to access，even with vpn.

2

This usually only happens when you don’t have enough GPU memory. Maybe you were running very close to the limit already?

How much actual free gpu memory do you have before you run the evaluation?

Oh, that is unfortunate. Can you give us more information about what isn’t working for you?

3

My GPU is nvidia p620 with 2GB RAM.
Maybe I should switch to CPU backend .

When I visit
https://community.konduit.ai
Most of the case I got timeout response

4

Yes, that seems like it might end up being not enough.

Oh, that is bad. I’m not sure how to help you there though :confused: It works fine for everyone who I’ve asked to test so far.