AgeCommit message (Expand)Author
22 hoursFix bug on CLCast from float to int8HEADmasterGiorgio Arena
22 hoursFix bug on CLGEMMMatrixMultiplyReshapedOnlyRHS testsGiorgio Arena
4 daysUpdate int8 A55 kernelsGeorgios Pinitas
5 daysUpdate errata with failures in CLCast and CLGEMMMatrixMultiplyReshapedOnlyRHS...Michele Di Giorgio
5 daysAdd proper range checking when identifying CPU capabilitiesGeorgios Pinitas
5 daysAvoid division with zeroGeorgios Pinitas
5 daysIntroduce Context opaque object of the new interfaceGeorgios Pinitas
6 daysAdd errata for failure in quantized CLFullyConnectedLayerMichele Di Giorgio
6 daysCheck cpu index is within a valid range when populating cpu informationGeorgios Pinitas
6 daysAccount Ksections correctly for interleaved and hybrid_indirect gemm kernelsGeorgios Pinitas
6 daysUse vector resizing instead of temporary assign in gemm_hybrid_indirectGeorgios Pinitas
6 daysAvoid division with zero on CLGEMM lhs/rhs configurationGeorgios Pinitas
7 daysFix redundant ternary operator in yolo node graph builderGeorgios Pinitas
7 daysRemove OpenGL ES supportManuel Bottini
7 daysAvoid nullptr dereference of vector_sum_colGeorgios Pinitas
7 daysRemove allocator check from BlobMemoryPool to avoid exceptionsGeorgios Pinitas
7 daysFix range checking of a value on a given data-type rangeGeorgios Pinitas
7 daysUse polymorphic_downcast instead of dynamic_castGeorgios Pinitas
7 daysRemove Compute Vision Neon supportMichalis Spyrou
7 daysRemove unreachable code in gemm_hybrid_quantized for computing k_blockGeorgios Pinitas
8 daysClose loaded library on OpenCL symbols destructionGeorgios Pinitas
8 daysEnable OpenMP scheduler support for clangGeorgios Pinitas
8 daysUpdate release notes with performance regression on G77 and updated toolchainGiorgio Arena
8 daysRemove 'actual's from building documentationGiorgio Arena
8 daysComply with Trademark rules for use of Neon, Arm and MaliSheri Zhang
8 daysDisable strict overflow warning for tensor_elem_atGiorgio Arena
11 daysUpdate Release NotesGiorgio Arena
11 daysFree _gemm_output_3d's memory region after run()Giorgio Arena
11 daysFix tolerance for NEReduceMean on aarch64Giorgio Arena
12 daysSet CLDirectConvolutionLayerKernel NCHW _border_size to input paddingGiorgio Arena
12 daysSet _gemm_output_3d's padding as the output's in NEGEMMConvolutionLayerGiorgio Arena
13 daysUpdate SONAME_VERSIONGiorgio Arena
13 daysUpdate deprecation noticesGiorgio Arena
13 daysUpdate LICENCEGiorgio Arena
13 daysAdd gitignore fileMichele Di Giorgio
13 daysRegenerate kernels and update A55 versionsGeorgios Pinitas
13 daysRedirect issues to Github onlyMichele Di Giorgio
14 daysHandle Conv2d layer with implicit output padding in NHWCGeorgios Pinitas
2021-02-15Fix data layout retention and handling of leftovers when there is no paddingMichele Di Giorgio
2021-02-15Adding Android S optionTeresa Charlin
2021-02-15Add WallClockTimer support for bare metal.Michalis Spyrou
2021-02-15Provide utlity function to print ITensorInfoMichele Di Giorgio
2021-02-15Add DIV to Elementwise layer support at graph levelAlessandro Navone
2021-02-11Building android-nn-driver failedManuel Bottini
2021-02-11Update OpenCL GEMM benchmark scriptGian Marco Iodice
2021-02-11Validate mlgo gemm type selection and fall back to default heuristicsSiCong Li
2021-02-10Fix tolerance issue in CL Winograd InputTransform FP16Manuel Bottini
2021-02-10Revert changes on tensor's strides and fix CLDepthwiseConvolution 3x3 QuantizedGiorgio Arena
2021-02-10Integrate MLGO into CLGEMMLowpMatrixMultiplyCore for native kernelSiCong Li
2021-02-10Decouple CpuPoolingKernel data type and data layoutSheri Zhang