diff options
Diffstat (limited to 'docs')
-rw-r--r-- | docs/00_introduction.dox | 68 | ||||
-rw-r--r-- | docs/01_library.dox | 10 | ||||
-rw-r--r-- | docs/06_functions_list.dox | 44 |
3 files changed, 59 insertions, 63 deletions
diff --git a/docs/00_introduction.dox b/docs/00_introduction.dox index 5c6b202139..3b340ebe5e 100644 --- a/docs/00_introduction.dox +++ b/docs/00_introduction.dox @@ -86,6 +86,50 @@ If there is more than one release in a month then an extra sequential number is @subsection S2_2_changelog Changelog +v21.05 Public major release + - Removed computer vision support from Neon backend + - Removed the following functions: + - NEAbsoluteDifference + - NEAccumulate + - NEBox3x3 + - NECannyEdge + - NEChannelCombine + - NEChannelExtract + - NEColorConvert + - NEDerivative + - NEDilate + - NEEqualizeHistogram + - NEErode + - NEFastCorners + - NEGaussian3x3 + - NEGaussian5x5 + - NEGaussianPyramid + - NEHOGDescriptor + - NEHOGDetector + - NEHOGGradient + - NEHOGMultiDetection + - NEHarrisCorners + - NEHistogram + - NEIntegralImage + - NELaplacianPyramid + - NELaplacianReconstruct + - NEMagnitude + - NEMeanStdDev + - NEMedian3x3 + - NEMinMaxLocation + - NENonLinearFilter + - NEOpticalFlow + - NEPhase + - NERemap + - NEScharr3x3 + - NESobel3x3 + - NESobel5x5 + - NESobel7x7 + - NETableLookup + - NEThreshold + - NEWarpAffine + - NEWarpPerspective + v21.02 Public major release - Various bug fixes. - Various optimisations. @@ -169,7 +213,7 @@ v20.11 Public major release - Removed padding from Neon kernels: - @ref NEComplexPixelWiseMultiplicationKernel - @ref NENonMaximaSuppression3x3Kernel - - @ref NERemapKernel + - NERemapKernel - @ref NEGEMMInterleave4x4Kernel - @ref NEDirectConvolutionLayerKernel - @ref NEScaleKernel @@ -462,7 +506,7 @@ v20.08 Public major release - Removed padding from: - @ref NEPixelWiseMultiplicationKernel - NEHeightConcatenateLayerKernel - - @ref NEThresholdKernel + - NEThresholdKernel - NEBatchConcatenateLayerKernel - @ref NETransposeKernel - @ref NEBatchNormalizationLayerKernel @@ -490,7 +534,7 @@ v20.08 Public major release - NEGEMMLowpQuantizeDownInt32ToUint8Scale - NEGEMMMatrixAccumulateBiasesKernel - Deprecated functions / interfaces: - - Non-descriptor based interfaces for @ref NEThreshold, @ref CLThreshold + - Non-descriptor based interfaces for NEThreshold, @ref CLThreshold - Non-descriptor based interfaces for @ref NEScale, @ref CLScale and @ref GCScale - In @ref NESoftmaxLayer, @ref NELogSoftmaxLayer, @ref CLSoftmaxLayer, @ref CLLogSoftmaxLayer and @ref GCSoftmaxLayer : The default "axis" value for @ref CLSoftmaxLayer, @ref CLLogSoftmaxLayer and @ref GCSoftmaxLayer is changed from 1 to 0. @@ -1270,13 +1314,13 @@ v17.04 Public bug fixes release - @ref CLGaussianPyramidHorKernel - @ref CLGaussianPyramidVertKernel - @ref CLGradientKernel - - @ref NEChannelCombineKernel + - NEChannelCombineKernel - @ref NEFillArrayKernel - - @ref NEGaussianPyramidHorKernel - - @ref NEGaussianPyramidVertKernel + - NEGaussianPyramidHorKernel + - NEGaussianPyramidVertKernel - NEHarrisScoreFP16Kernel - - @ref NEHarrisScoreKernel - - @ref NEHOGDetectorKernel + - NEHarrisScoreKernel + - NEHOGDetectorKernel - NELogits1DMaxKernel - NELogits1DShiftExpSumKernel - NELogits1DNormKernel @@ -1321,8 +1365,8 @@ v17.02.1 Sources preview - @ref CLMinMaxKernel, @ref CLMinMaxLocationKernel / @ref CLMinMaxLocation - @ref CLNonLinearFilterKernel / @ref CLNonLinearFilter - New Neon FP16 kernels (Requires armv8.2 CPU) - - @ref NEAccumulateWeightedFP16Kernel - - @ref NEBox3x3FP16Kernel + - NEAccumulateWeightedFP16Kernel + - NEBox3x3FP16Kernel - @ref NENonMaximaSuppression3x3FP16Kernel v17.02 Sources preview @@ -1333,8 +1377,8 @@ v17.02 Sources preview - @ref CLFastCornersKernel / @ref CLFastCorners - @ref CLMeanStdDevKernel / @ref CLMeanStdDev - New Neon kernels / functions: - - HOG / SVM: @ref NEHOGOrientationBinningKernel, @ref NEHOGBlockNormalizationKernel, @ref NEHOGDetectorKernel, NEHOGNonMaximaSuppressionKernel / @ref NEHOGDescriptor, @ref NEHOGDetector, @ref NEHOGGradient, @ref NEHOGMultiDetection - - @ref NENonLinearFilterKernel / @ref NENonLinearFilter + - HOG / SVM: NEHOGOrientationBinningKernel, NEHOGBlockNormalizationKernel, NEHOGDetectorKernel, NEHOGNonMaximaSuppressionKernel / NEHOGDescriptor, NEHOGDetector, NEHOGGradient, NEHOGMultiDetection + - NENonLinearFilterKernel / NENonLinearFilter - Introduced a CLScheduler to manage the default context and command queue used by the runtime library and create synchronisation events. - Switched all the kernels / functions to use tensors instead of images. - Updated documentation to include instructions to build the library from sources. diff --git a/docs/01_library.dox b/docs/01_library.dox index 848b060e9f..641fc3e11b 100644 --- a/docs/01_library.dox +++ b/docs/01_library.dox @@ -185,13 +185,13 @@ This is a very basic implementation which was originally used in the Neon runtim @sa CPPScheduler -@note Some kernels like for example @ref NEHistogramKernel need some local temporary buffer to perform their calculations. In order to avoid memory corruption between threads, the local buffer must be of size: ```memory_needed_per_thread * num_threads``` and a unique thread_id between 0 and num_threads must be assigned to the @ref ThreadInfo object passed to the ```run``` function. +@note Some kernels need some local temporary buffer to perform their calculations. In order to avoid memory corruption between threads, the local buffer must be of size: ```memory_needed_per_thread * num_threads``` and a unique thread_id between 0 and num_threads must be assigned to the @ref ThreadInfo object passed to the ```run``` function. @subsection S4_2_4 Functions Functions will automatically allocate the temporary buffers mentioned above, and will automatically multi-thread kernels' executions using the very basic scheduler described in the previous section. -Simple functions only call a single kernel (e.g @ref NEConvolution3x3), while more complex ones consist of several kernels pipelined together (e.g @ref NEGaussianPyramid, @ref NEHarrisCorners). Check their documentation to find out which kernels are used by each function. +Simple functions only call a single kernel (e.g @ref NEConvolution3x3), while more complex ones consist of several kernels pipelined together (e.g @ref NEFullyConnectedLayer ). Check their documentation to find out which kernels are used by each function. @code{.cpp} //Create a function object: @@ -230,11 +230,7 @@ For example: @subsection S4_4_2_cl_neon OpenCL / Neon interoperability -You can mix OpenCL and Neon kernels and functions. However it is the user's responsibility to handle the mapping/unmapping of OpenCL objects, for example: - -@snippet neoncl_scale_median_gaussian.cpp Neon / OpenCL Interop - -@sa main_neoncl_scale_median_gaussian +You can mix OpenCL and Neon kernels and functions. However it is the user's responsibility to handle the mapping/unmapping of OpenCL objects. @section S4_5_algorithms Algorithms diff --git a/docs/06_functions_list.dox b/docs/06_functions_list.dox index 96dce94a89..19485a8ac1 100644 --- a/docs/06_functions_list.dox +++ b/docs/06_functions_list.dox @@ -33,58 +33,39 @@ namespace arm_compute - @ref IFunction - @ref INESimpleFunction - - @ref NEAbsoluteDifference - @ref NEArithmeticAddition - @ref NEArithmeticSubtraction - @ref NEBoundingBoxTransform - - @ref NEBox3x3 - @ref NECast - @ref NEComplexPixelWiseMultiplication - @ref NEConvolution3x3 - @ref NEConvolutionRectangle - - @ref NEDilate - @ref NEElementwiseComparison - @ref NEElementwiseComparisonStatic - @ref NEElementwiseDivision - @ref NEElementwiseMax - @ref NEElementwiseMin - @ref NEElementwiseSquaredDiff - - @ref NEErode - @ref NEExpLayer - - @ref NEGaussian3x3 - - @ref NEIntegralImage - @ref NELogicalAnd - @ref NELogicalNot - @ref NELogicalOr - - @ref NEMedian3x3 - - @ref NENonLinearFilter - @ref NENonMaximaSuppression3x3 - @ref NEPixelWiseMultiplication - @ref NEPReluLayer - - @ref NERemap - @ref NEROIAlignLayer - @ref NERoundLayer - @ref NERsqrtLayer - - @ref NEScharr3x3 - @ref NESelect - - @ref NESobel3x3 - @ref NEStridedSlice - - @ref NEWarpAffine - - @ref NEWarpPerspective - @ref INESimpleFunctionNoBorder - - @ref NEAccumulate - - @ref NEAccumulateSquared - - @ref NEAccumulateWeighted - @ref NEActivationLayer - @ref NEBatchToSpaceLayer - @ref NEBitwiseAnd - @ref NEBitwiseNot - @ref NEBitwiseOr - @ref NEBitwiseXor - - @ref NEChannelCombine - - @ref NEChannelExtract - @ref NEChannelShuffleLayer - - @ref NEColorConvert - @ref NECopy - @ref NEDepthConvertLayer - @ref NEFlattenLayer @@ -93,23 +74,17 @@ namespace arm_compute - @ref NEGather - @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint - @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint - - @ref NEHOGDetector - - @ref NEMagnitude - @ref NEMeanStdDevNormalizationLayer - @ref NEPermute - - @ref NEPhase - @ref NEPriorBoxLayer - @ref NEReorgLayer - @ref NEReshapeLayer - @ref NEReverse - @ref NESlice - - @ref NETableLookup - - @ref NEThreshold - @ref NETile - @ref NETranspose - @ref NEArgMinMaxLayer - @ref NEBatchNormalizationLayer - - @ref NECannyEdge - @ref NEComplexPixelWiseMultiplication - @ref NEConcatenateLayer - @ref NEConvertFullyConnectedWeights @@ -121,11 +96,8 @@ namespace arm_compute - @ref NEDepthwiseConvolutionAssemblyDispatch - @ref NEDepthwiseConvolutionLayer - @ref NEDequantizationLayer - - @ref NEDerivative - @ref NEDetectionPostProcessLayer - @ref NEDirectConvolutionLayer - - @ref NEEqualizeHistogram - - @ref NEFastCorners - @ref NEFFT1D - @ref NEFFT2D - @ref NEFFTConvolutionLayer @@ -133,33 +105,19 @@ namespace arm_compute - @ref NEFillBorder - @ref NEFullyConnectedLayer - @ref NEFuseBatchNormalization - - @ref NEGaussian5x5 - - @ref NEGaussianPyramid - - @ref NEGaussianPyramidHalf - - @ref NEGaussianPyramidOrb - @ref NEGEMM - @ref NEGEMMAssemblyDispatch - @ref NEGEMMConv2d - @ref NEGEMMConvolutionLayer - @ref NEGEMMLowpMatrixMultiplyCore - @ref NEGenerateProposalsLayer - - @ref NEHarrisCorners - - @ref NEHistogram - - @ref NEHOGDescriptor - - @ref NEHOGGradient - - @ref NEHOGMultiDetection - @ref NEInstanceNormalizationLayer - @ref NEL2NormalizeLayer - - @ref NELaplacianPyramid - - @ref NELaplacianReconstruct - @ref NELSTMLayer - @ref NELSTMLayerQuantized - @ref NEQLSTMLayer - @ref NEMaxUnpoolingLayer - - @ref NEMeanStdDev - - @ref NEMinMaxLocation - @ref NENormalizationLayer - - @ref NEOpticalFlow - @ref NEPadLayer - @ref NEPoolingLayer - @ref NEQuantizationLayer @@ -169,8 +127,6 @@ namespace arm_compute - @ref NERNNLayer - @ref NEROIPoolingLayer - @ref NEScale - - @ref NESobel5x5 - - @ref NESobel7x7 - @ref NESoftmaxLayerGeneric <IS_LOG> - @ref NESpaceToBatchLayer - @ref NESpaceToDepthLayer |