- ba2dd82 Benchmark and set default LWS for GEMM, Direct convolution and Winograd by Giorgio Arena · 3 years ago
- d87aded Port CLGEMMConvolutionLayer by Manuel Bottini · 3 years, 1 month ago
- 7075fe2 Reorganize the kernels into nhwc, nchw and common folders by Adnan AlSinan · 3 years, 1 month ago
- 767dbf9 Fix oclgrind int overflow warning by Freddie Liardet · 3 years ago
- d844c08 Port CLIm2ColKernel to ClIm2ColKernel by Manuel Bottini · 3 years, 1 month ago
- 7b23732 Port CLCol2ImKernel to ClCol2ImKernel by Manuel Bottini · 3 years, 1 month ago
- c38ca38 Fix CL kernel compilation failure by Michalis Spyrou · 3 years, 1 month ago
- a387e27 Add in-place calculation support for CL elementwise arithmetic kernels by Sheri Zhang · 3 years, 1 month ago
- 04b39e8 Limit the LOOP_UNROLLING on the kernel height by Gian Marco Iodice · 3 years, 1 month ago
- da81675 Remove redundant implementations of Add/Sub operators by Georgios Pinitas · 3 years, 1 month ago
- 5e28181 Fix manual LOOP_UNROLLING by Gian Marco Iodice · 3 years, 1 month ago
- 8155c02 Rework OpenCL Depthwise Convolution by Gian Marco Iodice · 3 years, 4 months ago
- bc4e311 Implement FP GPU depthwise convolution 1x1 kernel for in-place computation by SiCongLi · 3 years, 1 month ago
- 6683165 Add quantization helper functions for OpenCL by Georgios Pinitas · 3 years, 1 month ago
- f7f6450 Adjust minimum DDK version for manual unroll by Giorgio Arena · 3 years, 1 month ago
- c63b722 Revert "Rework OpenCL Depthwise Convolution" by Gian Marco Iodice · 3 years, 1 month ago
- 4a578b9 Port the ClGemmLowp kernels to the new API by Georgios Pinitas · 3 years, 1 month ago
- 72b5687 Enable global pooling optimization on OpenCL by Gian Marco Iodice · 3 years, 1 month ago
- 4a95bba Set up the framework to choose the default LWS by Giorgio Arena · 3 years, 1 month ago
- 5f6fdc1 Rework gemmlowp reshaped_only_rhs using the new macros by Giorgio Arena · 3 years, 2 months ago
- 561c176 Rework OpenCL Depthwise Convolution by Gian Marco Iodice · 3 years, 4 months ago
- ef5aac6 Add FP16 support to CLRemap by Freddie Liardet · 3 years, 2 months ago
- c6f4ec3 Port CLWinogradConvolutionLayer with ClWinogradConv2d by Manuel Bottini · 3 years, 2 months ago
- 36dff9f Add NHWC support to CLRemap by Frederick Liardet · 3 years, 3 months ago
- 1883493 Revert "Add optimization for global pooling in pooling_layer.cl" by Pablo Tello · 3 years, 2 months ago
- 2098445 Fix bug in PReluLayer when input is 1xN size by Freddie Liardet · 3 years, 3 months ago
- 9fc3be6 Fuse activation in ClDirectConv2dKernel for float types by Georgios Pinitas · 3 years, 2 months ago
- b4bb6a0 Rename ported functions by Manuel Bottini · 3 years, 2 months ago
- bf1dbd8 Fix retrieving device architecture twice in CLCompileContext by Giorgio Arena · 3 years, 2 months ago
- ea8d266 Enable unroll through pragma based on DDK version by Giorgio Arena · 3 years, 2 months ago
- b7b6066 clCreateKernel failure of CL/ChannelShuffle/U8/ by Manuel Bottini · 3 years, 2 months ago
- 11d8415 Port DepthConvert to new Api by Georgios Pinitas · 3 years, 3 months ago
- 856f66e Port CLGEMM to memory injecting interface by Georgios Pinitas · 3 years, 3 months ago
- bdd16d1 Add macro to manually unroll loops in OpenCL by Giorgio Arena · 3 years, 3 months ago
- 538a076 Fix oclgrind error on CLGEMMLowp reshaped only RHS quantized per channel by Giorgio Arena · 3 years, 3 months ago
- ed4b8a0 Fix MeanStdDevNormalizationLayer reference outputting nan for FP16 by Giorgio Arena · 3 years, 3 months ago
- 7f8caf7 Fix integer overflow and null dereference by Michalis Spyrou · 3 years, 3 months ago
- e57e11d Remove padding from CLChannelShuffleLayerKernel by Manuel Bottini · 3 years, 3 months ago
- fd83bc8 Fix GEMMLowp output stage validation crash when input's first dimension == 1 by Giorgio Arena · 3 years, 3 months ago
- c2d0e20 Fix bug in Select operator when input is 1xN by Freddie Liardet · 3 years, 3 months ago
- 7d2f69f Fix 'ARM_DOT_K0XN0' macro redefined by Giorgio Arena · 3 years, 3 months ago
- 2cd5b31 Remove unused CLCoreRuntimeContext by Georgios Pinitas · 3 years, 3 months ago
- 2ba39b6 Fix missing DATA_TYPE in DOT_PRODUCT4_INTEGER8 OpenCL macro by Gian Marco Iodice · 3 years, 3 months ago
- 78e7691 Remove TODOs by Sheri Zhang · 3 years, 3 months ago
- 908f616 Creates ClKerneLibrary to serve just as a kernel container without build logic by Georgios Pinitas · 3 years, 3 months ago
- 448cb45 Adding S32 support to CLPixelWiseMultiplication by Suhail Munshi · 3 years, 3 months ago
- 1fac037 Fix bug on CLReductionOperation by Giorgio Arena · 3 years, 3 months ago
- 6268742 Update operator list documentation. Part 2. by Teresa Charlin · 3 years, 3 months ago
- 50929ef Add optimization for global pooling in pooling_layer.cl by Gian Marco Iodice · 3 years, 3 months ago
- baebdb9 Remove Global pooling optimization by Michalis Spyrou · 3 years, 3 months ago
- 3ecf9fe Remove OpenCL padding: CLReductionOperationKernel by Giorgio Arena · 3 years, 3 months ago
- 223aba9 Remove stale/solved TODOs by Michele Di Giorgio · 3 years, 3 months ago
- 9f6111a Fix Global Pooling failures by Michalis Spyrou · 3 years, 3 months ago
- c3c352e Add Queue support by Georgios Pinitas · 3 years, 5 months ago
- 2b7fee0 Add explicit cast on GEMMLowp kernel by Michalis Spyrou · 3 years, 3 months ago
- 40471d1 Add optimization for global pooling in pooling_layer.cl by Gian Marco Iodice · 3 years, 3 months ago
- 3eb5d29 Remove padding from CLROIPoolingLayerKernel by Gian Marco Iodice · 3 years, 3 months ago
- 5c3eeec Fixed CTS failures CLInstanceNorm by Pablo Tello · 3 years, 3 months ago
- 6717688 [Nightly #1129] CL/Winograd/ConvolutionLayer/F16 mismatch on Mate9 by Manuel Bottini · 3 years, 3 months ago
- dd2d013 Fix bug on CLPixelWiseMultiplication broadcasted for Quantized types by Giorgio Arena · 3 years, 3 months ago
- 402740d Add support for CLVK by Michalis Spyrou · 3 years, 3 months ago
- bff2f9f Remove unused yolo_layer OpenCL kernel by Michele Di Giorgio · 3 years, 3 months ago
- dcf4c87 CLDepthwiseConvolutionLayer rework - Part 1 by Giorgio Arena · 3 years, 4 months ago
- ada6cbc Remove OpenCL padding: CLPixelWiseMultiplicationKernel by Giorgio Arena · 3 years, 4 months ago
- 031d6a9 [Nightly #1129] CL/Winograd/ConvolutionLayer/F16 mismatch on Mate9 by Manuel Bottini · 3 years, 3 months ago
- b25883a Remove experimental tracing feature by Pablo Marquez Tello · 3 years, 3 months ago
- ae1c9fe Added S32 Integer support to DIV operator in CLElementWiseOperations with Tests by Suhail Munshi · 3 years, 4 months ago
- fe7ae81 CLInstanceNormalizationLayer NHWC optimisation by Pablo Marquez Tello · 3 years, 5 months ago
- 60c3b0e Port DepthwiseConvolution to new API by Michalis Spyrou · 3 years, 4 months ago
- 4f1650f Remove padding from CLNormalizePlanarYUVLayerKernel by Sheri Zhang · 3 years, 4 months ago
- d05f67f Remove unnecessary use AccessWindows by Michele Di Giorgio · 3 years, 4 months ago
- 3795491 Add GEMM heuristic for Mali-G78 by Gian Marco Iodice · 3 years, 4 months ago
- 142e4ca Remove unused AccessWindow* includes by Michele Di Giorgio · 3 years, 4 months ago
- 21c2895 Remove OpenCL padding: CLNormalizationLayerKernel by Manuel Bottini · 3 years, 4 months ago
- 91b7f74 Port CLConvertFullyConnectedWeights to new API by Teresa Charlin · 3 years, 4 months ago
- 598e3a8 Fix TILE initialization in direct convolution and winograd transforms by Gian Marco Iodice · 3 years, 4 months ago
- 21fda6f Fix T_LOAD too few parameters issue by Sheri Zhang · 3 years, 4 months ago
- 0b76f7d Add support for cl_image in CLDirectConvolutionLayer by Gian Marco Iodice · 3 years, 4 months ago
- 6dbcc0e Fix OpenCL kernel compiling failure with array initializer by Sheri Zhang · 3 years, 4 months ago
- 99b1a1c Fix bug on Implicit Padding for CL GEMMMatrixMultiplyInterleavedTransposed by Manuel Bottini · 3 years, 4 months ago
- 83eee19 Winograd Output transform 7x7 rework by Giorgio Arena · 3 years, 4 months ago
- ec82b95 Fix OpenCL kernel compiling failure with array initilizer by Sheri Zhang · 3 years, 4 months ago
- 8778108 Rewritten Winograd (4x4, 5x5) output transformation by Aleksandr Nikolaev · 3 years, 4 months ago
- 4a5f73d Ensure OpenCL runtimes are initialized first by Marco Antognini · 3 years, 4 months ago
- 534b889 Rework the OpenCL Winograd Input Transformations NHWC by Gian Marco Iodice · 3 years, 4 months ago
- 7a452fe Remove OpenCL padding: CLL2NormalizeLayerKernel by Manuel Bottini · 3 years, 4 months ago
- 4ed7b39 Added Qasymm8 datatype support to CLROIPoolingLayer with Tests by Suhail Munshi · 3 years, 4 months ago
- 33f41fa Fix trademarks throughout the codebase by Michele Di Giorgio · 3 years, 5 months ago
- c6f9510 Remove Computer Vision generic interfaces and types by Georgios Pinitas · 3 years, 4 months ago
- 2788609 Port ClTranspose to new API by Teresa Charlin · 3 years, 5 months ago
- 702dc0c Remove usage of valid window region CL - NHWC by Michalis Spyrou · 3 years, 4 months ago
- 2ca5b08 New variant of OpenCL Winograd (4x4,5x5) input transformation by Aleksandr Nikolaev · 3 years, 5 months ago
- 97e2580 Check biases pointer before referencing in CLDirectConvolutionLayer by Michele Di Giorgio · 3 years, 4 months ago
- a8903c8 Improve performance of Winograd Output Transform 3x3 by Gian Marco Iodice · 3 years, 4 months ago
- b5323cf Fix gemmlowp kernel crash when n0==16 by SiCongLi · 3 years, 5 months ago
- 049989a Improve performance of Winograd Input transform 3x3 by Giorgio Arena · 3 years, 4 months ago
- 5c9eed8 Extend direct convolution (F32/F16/QASYMM8) by Gian Marco Iodice · 3 years, 4 months ago
- 1efed92 Make ClDirectConvolutionKernel stateless by Sheri Zhang · 3 years, 5 months ago
- f9ab9f9 Make ClPixelWiseMultiplicationKernel stateless by Sheri Zhang · 3 years, 5 months ago
- 42bd265 Revert "Close loaded library on OpenCL symbols destruction" by Georgios Pinitas · 3 years, 5 months ago