Package deepboof.impl.backward.standard
-
Class Summary Class Description BaseDBatchNorm_F64 Implements common functionality for all batch normalization implementations forTensor_F64.BaseDFunction<T extends deepboof.Tensor<T>> Base class which implements common functionality between allDFunctionDActivationReLU_F64 Implementation ofDActivationReLUforTensor_F64DActivationSigmoid_F64 Implementation ofDActivationSigmoidforTensor_F64.DActivationTanH_F64 Implementation ofDActivationTanHforTensor_F64.DClippedPadding2D_F64 Backwards implementation ofClippedPadding2D_F64.DConstantPadding2D_F64 Backwards implementation ofConstantPadding2D_F64.DFunctionBatchNorm_F64 Implementation ofDFunctionBatchNormforTensor_F64.DFunctionDropOut_F64 Implementation ofDFunctionDropOutforTensor_F64DFunctionLinear_F64 Implementation ofDFunctionLinearforTensor_F64DSpatialBatchNorm_F64 Implementation ofDSpatialBatchNormforTensor_F64.DSpatialConvolve2D_F64 Implementation ofDSpatialConvolve2DforTensor_F64.DSpatialMaxPooling_F64 DSpatialWindowChannel<T extends deepboof.Tensor<T>,P extends DSpatialPadding2D<T>> Backwards functions for operations which convolve a window across the input spatial tensor and process the image in a BCHW (batch, channel, (row, column)) order, e.g.DSpatialWindowImage<T extends deepboof.Tensor<T>,P extends DSpatialPadding2D<T>> Backwards functions for operations which convolve a window across the input spatial tensor.ElementWiseDFunction<T extends deepboof.Tensor<T>> Base class for element-wise derivative functionsNumericalGradient_F64 Implementation ofNumericalGradientforTensor_F64