aboutsummaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Expand)Author
2018-11-02COMPMID-582: Add validation to channel_extract kernels.Ioan-Cristian Szabo
2018-11-02COMPMID-882 - Optimizing GEMMLowp on OpenCL reshaping matricesGian Marco
2018-11-02COMPMID-905 Optimize CLSoftmaxLayer for QASYMM8Giorgio Arena
2018-11-02COMPMID-765 Move direct convolution output stage to the right fileGiorgio Arena
2018-11-02COMPMID-856: CL Depthwise Convolution QASYMM8 supportGeorgios Pinitas
2018-11-02COMPMID-906: Use fused activation in NEON Batch normalizationGeorgios Pinitas
2018-11-02COMPMID-765: Fix CPPPermute error when permuting the strides.Georgios Pinitas
2018-11-02COMPMID-765 : NEON Wrapper initial traits and overloadsGeorgios Pinitas
2018-11-02COMPMID-765: Switch 1x1 DeconvolutionLayer to use the ConvolutionLayerGeorgios Pinitas
2018-11-02COMPMID-765: Sanitize permutation vector for Permute.Georgios Pinitas
2018-11-02COMPMID-905 Asymm functions support for all vec sizesGiorgio Arena
2018-11-02COMPMID-903: Implements NEPermute for NHWC conversionsGeorgios Pinitas
2018-11-02COMPMID-765: Fix CLDeconvolutionLayerUpsampleKernel access window.Georgios Pinitas
2018-11-02COMPMID-875: Deconvolution 4x4 not workingGeorgios Pinitas
2018-11-02COMPMID-884: Valgrind: NEDirectConvolutionLayerKernel invalid readGeorgios Pinitas
2018-11-02COMPMID-828 - Add support for pool widths 4, 5 & 6 and for non square data si...Isabella Gottardi
2018-11-02COMPMID-765 - Extended GEMM benchmarkGian Marco
2018-11-02COMPMID-897 Merge batch normalization with bounded reluGiorgio Arena
2018-11-02COMPMID-892: OCLGrind failures on both validation and benchmarkGeorgios Pinitas
2018-11-02COMPMID-907 Optimizing FixedPoint calculation in the output stage of GEMMLowpGiorgio Arena
2018-11-02COMPMID-765: Fixes DepthwiseConvolution weights shapeGeorgios Pinitas
2018-11-02COMPMID-765: Fix inclusion error.Georgios Pinitas
2018-11-02COMPMID-873: Integrate RSH NEON Depthwise Convolution routineGeorgios Pinitas
2018-11-02COMPMID-876: Integrate RSH native GEMM kernel.Georgios Pinitas
2018-11-02COMPMID-578: Implement FAST corners for CL/NEONAbe Mbise
2018-11-02COMPMID-888 Valgrind invalid read in NEGEMVAArch64KernelMichalis Spyrou
2018-11-02COMPMID-901 - Optimizing CLCol2ImKernelGian Marco
2018-11-02COMPMID-879: Investigate CL mismatches in Convolution S16Georgios Pinitas
2018-11-02COMPMID-828 - Add support for non square pool size - Part1Isabella Gottardi
2018-11-02COMPMID-895 - Optimizing CLDepthwiseConvolution3x3KernelGian Marco
2018-11-02COMPMID-894: Segfault: neon_cnn on S5 neoGeorgios Pinitas
2018-11-02COMPMID-765 - Added LWS hint in CLIm2ColGian Marco
2018-11-02COMPMID-784: Added support for PADDING = SAME in Winograd layer.Pablo Tello
2018-11-02COMPMID-891 - Use OpenCL timer in CLTunerGian Marco
2018-11-02COMPMID-855 - Optimizing im2col on OpenCL (DCHW)Gian Marco
2018-11-02COMPMID-754: Add validation method to CLPermute kernelMichele Di Giorgio
2018-11-02COMPMID-874: Improve default number of threads choice in the SchedulerGeorgios Pinitas
2018-11-02COMPMID-890: Valgrind: NEON Convolution Layer validation failsIsabella Gottardi
2018-11-02COMPMID-871: Remove vst4q/vld4q from NEActivationLayer.Georgios Pinitas
2018-11-02COMPMID-787: Add CL support for broadcast multiplyMichele Di Giorgio
2018-11-02COMPMID-765: Fix direct convolution output stage.Georgios Pinitas
2018-11-02COMPMID-765: Fixed clangtidy warnings.Pablo Tello
2018-11-02COMPMID-765: Fixed CPU detection code to read larger buffer (was failing on A...Pablo Tello
2018-11-02COMPMID-784: Winograd tramsforms refactoringPablo Tello
2018-11-02APPBROWSER-400: Implement the tensorshift kernel for fixing DC's alignment is...Xinghang Zhou
2018-11-02APPBROWSER-394: GLES fails to compile Dropout kernel on S5steli01
2018-11-02IVGCVSW-863 Broadcast support in CL/NEON Arithmetic AddDiego Lopez Recas
2018-11-02COMPMID-866: Integrate SGEMV Neon Assembly from RSHMichele Di Giorgio
2018-11-02COMPMID-765: Allow RSH's code to not have default cases in their switchesAnthony Barbier
2018-11-02APPBROWSER-390,397,398: bugfix and fully connected validation issue on specif...zhenglin