• Example Search
• Project Search

• Java
• Python
• JavaScript
• TypeScript
• C++
• Scala
• Blog

• cct-nn-master
  • src
    • main
      • scala
        • toolkit
          • neuralnetwork
            • source
              • ByteFilePackedSensor.scala
              • FloatDataSource.scala
              • FloatLabelSource.scala
              • StreamLoader.scala
              • RandomSource.scala
              • SaveParameters.scala
              • ByteLabelSource.scala
              • FloatFileSensor.scala
              • ByteFileSensor.scala
              • ByteDataSource.scala
            • WeightStore.scala
            • function
              • Convolution.scala
              • CrossEntropySoftmax.scala
              • StackFeatures.scala
              • FrequencyConvolution.scala
              • Subspace.scala
              • Softmax.scala
              • BasicOps.scala
              • FullyConnected.scala
              • Tanh.scala
              • MaxPooling.scala
              • Downsample.scala
              • AplusBXtoN.scala
              • RightPad.scala
              • CrossEntropySoftmaxes.scala
              • CrossEntropy.scala
              • ReLU.scala
              • TrainableState.scala
              • SumOfSquares.scala
              • Dropout.scala
              • Bias.scala
              • SpatialConvolution.scala
              • ZeroPad.scala
              • RightCrop.scala
            • policy
              • StandardLearningRule.scala
              • ConvolutionalLayerPolicy.scala
              • GaussianInit.scala
              • Distribution.scala
              • EmptyBinding.scala
              • ZeroInit.scala
              • StubInit.scala
              • ConvInit.scala
              • WeightInitPolicy.scala
              • FCInit.scala
              • LearningRule.scala
            • DifferentiableField.scala
            • layer
              • ConvolutionLayer.scala
              • BiasLayer.scala
              • Layer.scala
              • FullyConnectedLayer.scala
            • operator
              • spray.scala
              • fourierBackProject.scala
              • GPUOperatorHelper.scala
              • backFC.scala
              • borderReduceSum.scala
              • weightGradFC.scala
              • fourierFilterGradMAC.scala
              • fftReal.scala
              • indexToOneHotCode.scala
              • cropFilters.scala
              • fourierProjectMAC.scala
              • expandFilters.scala
              • fourierProject.scala
              • fourierFilterGrad.scala
              • fftInverseReal.scala
              • forwardFC.scala
              • fourierBackProjectMAC.scala
              • sumSpray.scala
            • util
              • CorrectCount.scala
              • LowPass.scala
              • NormalizedLowPass.scala
              • Norm.scala
              • Classify.scala
            • Implicits.scala
            • GradientPropagation.scala
            • WeightBinding.scala
            • DifferentiableFieldOps.scala
    • test
      • scala
        • toolkit
          • neuralnetwork
            • ComputeTests.scala
            • source
              • FloatFileSensorSpec.scala
              • ByteFileSensorSpec.scala
              • ByteFilePackedSensorSpec.scala
            • UnitSpec.scala
            • function
              • SumOfSquaresSpec.scala
              • DownsampleSpec.scala
              • AddConstantSpec.scala
              • AplusBXtoNSpec.scala
              • RightCropSpec.scala
              • BiasSpec.scala
              • TanhSpec.scala
              • RightPadSpec.scala
              • DivideConstantSpec.scala
              • SubspaceSpec.scala
              • DivideSpec.scala
              • CrossEntropySoftmaxSpec.scala
              • MultiplyConstantSpec.scala
              • FullyConnectedSpec.scala
              • AddSpec.scala
              • PowSpec.scala
              • MultiplySpec.scala
              • MaxPoolingFunctionalSpec.scala
              • ZeroPadSpec.scala
              • ReLUSpec.scala
              • DropoutSpec.scala
              • FrequencyConvolutionSpec.scala
              • SoftmaxSpec.scala
              • CrossEntropySoftmaxesSpec.scala
              • CrossEntropySpec.scala
              • MaxPoolingSpec.scala
              • StackFeaturesSpec.scala
              • SpatialConvolutionSpec.scala
            • examples
              • ComputeMean.scala
              • DualPortRegression.scala
              • CIFARTrainer.scala
              • networks
                • Net.scala
                • SimpleConvNet.scala
                • CIFAR.scala
                • AlexNet.scala
              • util
                • RandomPatch.scala
                • DataAugmentation.scala
                • AveragePooling.scala
                • VectorMeanSquaresBorderCyclic.scala
                • RandomFlip.scala
                • VectorMeanSquares.scala
                • PlanarConvolutionSpec.scala
                • PlanarConvolution.scala
              • AlexNetDebugger.scala
              • ConvNet.scala
              • LogisticRegression.scala
              • Derivative.scala
              • CIFARDebugger.scala
              • TwoLayerNet.scala
            • AdjointTestSource.scala
            • operator
              • fftInverseRealSpec.scala
              • CyclicFilterAdjointFFTDebugger.scala
              • ConvBackDebugger.scala
              • ConvLayerDebugger.scala
              • fftRealSpec.scala
              • ConvFwdDebugger.scala
              • ValidFilterAdjointFFTDebugger.scala
              • FCFwdDebugger.scala
              • BorderReduceSumDebugger.scala
              • ConvGradDebugger.scala
            • DifferentiableFieldSpec.scala
            • performance
              • FourierProjectMACTest.scala
              • AlexNetTest.scala
              • Benchmark.scala
  • project
    • build.properties
    • plugins.sbt
  • .travis.yml
  • README.md
  • build.sbt
  • .gitignore
  • LICENSE.txt

cct-nn

Deep learning library for the HPE Cognitive Computing Toolkit (CCT). The NN library layers a DifferentiableField type on top of regular CCT fields. A DifferentiableField behaves like a normal field in most contexts, but includes additional state that allow it to carry gradients. Operators on DifferentiableField types include both forward and backward functions. The NN toolkit can propagate gradients from a loss function back through a directed acyclic graph of differentiable fields and operators.

If you're new to CCT, start with the tutorial.