- 702dc0c Remove usage of valid window region CL - NHWC by Michalis Spyrou · 3 years, 4 months ago
- 26c5c0f NEGEMMConvolution and NEGEMMConv2d produce different results by Manuel Bottini · 3 years, 4 months ago
- 2ca5b08 New variant of OpenCL Winograd (4x4,5x5) input transformation by Aleksandr Nikolaev · 3 years, 4 months ago
- 1e3ab42 Make CpuPixelWiseMultiplicationKernel stateless by Sheri Zhang · 3 years, 4 months ago
- e81825b Workaround for compiler error in gcc-9.2 and 9.3 by Pablo Marquez Tello · 3 years, 4 months ago
- 97e2580 Check biases pointer before referencing in CLDirectConvolutionLayer by Michele Di Giorgio · 3 years, 4 months ago
- 40efd53 Remove GLES-related code 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
- 32e0646 Fix indexing of SVE ArithmeticAddition kernel by Sang-Hoon Park · 3 years, 4 months ago
- b5323cf Fix gemmlowp kernel crash when n0==16 by SiCongLi · 3 years, 5 months ago
- 383de02 Fixed compiler errors by Pablo Marquez Tello · 3 years, 4 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, 4 months ago
- f9ab9f9 Make ClPixelWiseMultiplicationKernel stateless by Sheri Zhang · 3 years, 4 months ago
- d1dc09c Port CpuTranspose to new API by Teresa Charlin · 3 years, 5 months ago
- 226169f Fix SVE build on newer compilers by Michalis Spyrou · 3 years, 4 months ago
- ab84088 Added Qasymm8 datatype support to NEROIPoolingLayer with Tests by Suhail Munshi · 3 years, 5 months ago
- a50f193 Updated cpu detection by Pablo Marquez Tello · 3 years, 4 months ago
- 10b3826 Port Arm(R) Neon(TM) Scale to new API by Manuel Bottini · 3 years, 5 months ago
- 42bd265 Revert "Close loaded library on OpenCL symbols destruction" by Georgios Pinitas · 3 years, 4 months ago
- 3b131ab Port OpenCL Scale to new API by Manuel Bottini · 3 years, 5 months ago
- 9e73c93 Port OpenCL Dequantization to new API by Manuel Bottini · 3 years, 5 months ago
- 6b0bf99 Add CpuPRelu operators by Sang-Hoon Park · 3 years, 5 months ago
- d7de9c5 Change the order of internal tensor allocation in CLSoftmaxLayer by Sang-Hoon Park · 3 years, 4 months ago
- 00f4dfc Port Arm(R) Neon(TM) Dequantization to new API by Manuel Bottini · 3 years, 4 months ago
- 5a1bf62 Port OpenCL Quantization to new API by Manuel Bottini · 3 years, 5 months ago
- 0ded4c4 Port Arm(R) Neon(TM) Quantization to new API by Manuel Bottini · 3 years, 4 months ago
- 86ee237 Use memalign on bare metal by Michalis Spyrou · 3 years, 4 months ago
- d0b7b4b Remove valid region calculation from CpuElementwise by Sang-Hoon Park · 3 years, 4 months ago
- 72f13bd Fix segmentation fault on CLSoftmaxLayer by Sang-Hoon Park · 3 years, 4 months ago
- 668ccdc Add dynamic tensor support to CpuElementwise by Sang-Hoon Park · 3 years, 5 months ago
- 201e0fe Make Softmax kernels on OpenCL stateless by Sang-Hoon Park · 3 years, 6 months ago
- 023efb3 Make auxiliary tensor within NEGEMMConvolutionLayer resizable by Georgios Pinitas · 3 years, 4 months ago
- b88272e Remove usage of valid window region in NHWC CPU kernels - Part2 by SiCongLi · 3 years, 5 months ago
- c7b1e84 Remove usage of valid window region in NHWC CPU kernels - Part1 by SiCongLi · 3 years, 5 months ago
- b861074 Fix build errors for functions requiring SVE2 extension by Sang-Hoon Park · 3 years, 4 months ago
- 57d7366 Add log on memory allocation failure in CpuContext by Sang-Hoon Park · 3 years, 5 months ago
- dcf3c7e Move utility functions to NE/SVEMath by Sang-Hoon Park · 3 years, 5 months ago
- 5ff38da Create ClPRelu operator by Sang-Hoon Park · 3 years, 5 months ago
- 473cb01 Remove Compute Vision CL support by Michalis Spyrou · 3 years, 5 months ago
- f466d75 Add QASYMM8_SIGNED support to graph examples via graph mutator by SiCongLi · 3 years, 5 months ago
- c1b9b09 Fix bug on CLCast from float to int8 by Giorgio Arena · 3 years, 5 months ago
- aac71e5 Update int8 A55 kernels by Georgios Pinitas · 3 years, 5 months ago
- 05784ad Add proper range checking when identifying CPU capabilities by Georgios Pinitas · 3 years, 5 months ago
- 9a67178 Avoid division with zero by Georgios Pinitas · 3 years, 5 months ago
- 8a5146f Introduce Context opaque object of the new interface by Georgios Pinitas · 3 years, 6 months ago
- 8c52c59 Check cpu index is within a valid range when populating cpu information by Georgios Pinitas · 3 years, 5 months ago
- 6f45cf7 Account Ksections correctly for interleaved and hybrid_indirect gemm kernels by Georgios Pinitas · 3 years, 5 months ago
- 85e16c2 Use vector resizing instead of temporary assign in gemm_hybrid_indirect by Georgios Pinitas · 3 years, 5 months ago
- 261df74 Avoid division with zero on CLGEMM lhs/rhs configuration by Georgios Pinitas · 3 years, 5 months ago
- 2ae0860 Fix redundant ternary operator in yolo node graph builder by Georgios Pinitas · 3 years, 5 months ago
- ceaa0bf Remove OpenGL ES support by Manuel Bottini · 3 years, 5 months ago
- c40562d Avoid nullptr dereference of vector_sum_col by Georgios Pinitas · 3 years, 5 months ago
- 6061a89 Remove allocator check from BlobMemoryPool to avoid exceptions by Georgios Pinitas · 3 years, 5 months ago
- b0cd5d8 Use polymorphic_downcast instead of dynamic_cast by Georgios Pinitas · 3 years, 5 months ago
- 27e67f0 Remove Compute Vision Neon support by Michalis Spyrou · 3 years, 5 months ago
- 0ad0129 Remove unreachable code in gemm_hybrid_quantized for computing k_block by Georgios Pinitas · 3 years, 5 months ago
- d62ef4d Close loaded library on OpenCL symbols destruction by Georgios Pinitas · 3 years, 5 months ago
- ac6499a Comply with Trademark rules for use of Neon, Arm and Mali by Sheri Zhang · 3 years, 5 months ago
- f55e3fb Free _gemm_output_3d's memory region after run() by Giorgio Arena · 3 years, 5 months ago
- 9a09180 Set CLDirectConvolutionLayerKernel NCHW _border_size to input padding by Giorgio Arena · 3 years, 5 months ago
- d820ed7 Set _gemm_output_3d's padding as the output's in NEGEMMConvolutionLayer by Giorgio Arena · 3 years, 5 months ago
- 5b50f42 Update deprecation notices by Giorgio Arena · 3 years, 5 months ago
- e28cf39 Regenerate kernels and update A55 versions by Georgios Pinitas · 3 years, 6 months ago
- 2ee9801 Handle Conv2d layer with implicit output padding in NHWC by Georgios Pinitas · 3 years, 5 months ago
- bae2237 Fix data layout retention and handling of leftovers when there is no padding by Michele Di Giorgio · 3 years, 5 months ago
- 6413e49 Add DIV to Elementwise layer support at graph level by Alessandro Navone · 3 years, 6 months ago
- 8958167 Building android-nn-driver failed by Manuel Bottini · 3 years, 5 months ago
- 1a28e73 Validate mlgo gemm type selection and fall back to default heuristics by SiCong Li · 3 years, 5 months ago
- edc524e Revert changes on tensor's strides and fix CLDepthwiseConvolution 3x3 Quantized by Giorgio Arena · 3 years, 5 months ago
- db35345 Integrate MLGO into CLGEMMLowpMatrixMultiplyCore for native kernel by SiCong Li · 3 years, 5 months ago
- 79144a6 Decouple CpuPoolingKernel data type and data layout by Sheri Zhang · 3 years, 5 months ago
- 8c23ba1 Integrate MLGO into CLGEMM for reshaped kernel by SiCong Li · 3 years, 5 months ago
- 373b407 Make Softmax kernels and operator stateless by Michalis Spyrou · 3 years, 6 months ago
- 4841c97 Add mlgo to graph examples by SiCong Li · 3 years, 5 months ago
- 9f7d55a Fix CLDepthwiseConvolutionLayer 3x3 QASYMM8 by Giorgio Arena · 3 years, 5 months ago
- bbd8fac Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part4 by SiCong Li · 3 years, 5 months ago
- bd8b1e2 Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part3 by SiCong Li · 3 years, 5 months ago
- db4a6c1 Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part2 by SiCong Li · 3 years, 5 months ago
- 70858d8 Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part1 by SiCong Li · 3 years, 5 months ago
- 8e21331 ArmNN internal OOB failed to build ACL with the use of a deleted function by Manuel Bottini · 3 years, 5 months ago
- 7e20e29 Make memset/copy functions state-less by Sheri Zhang · 3 years, 6 months ago
- 1928904 Make NEON Pooling kernels and functions state-less by Michele Di Giorgio · 3 years, 5 months ago
- cf87f50 Tweak scheduling use of SQDMULH in quantized AVG pooling by Michele Di Giorgio · 3 years, 6 months ago
- 76cb751 Fix CL Compiler frontend failure for ArgMinMax by Giorgio Arena · 3 years, 5 months ago
- 0841ca0 Fix OpenCL direct convolution by Gian Marco Iodice · 3 years, 6 months ago
- 8d8a1c5 Fix segfault in fsrcnn.tflite in GpuAcc by Teresa Charlin · 3 years, 5 months ago
- e131466 Make CL Pooling kernels and functions state-less by Michele Di Giorgio · 3 years, 6 months ago
- be9f9f9 Add WBSM tuning to CLTuner by Manuel Bottini · 3 years, 6 months ago
- 7061eb2 Implement MLGO module by SiCong Li · 3 years, 6 months ago
- 74a142c Account for eventual padding in Pooling assembly kernels by Michele Di Giorgio · 3 years, 6 months ago
- ea85727 Compress OpenCL kernel files using zlib for Android by Georgios Pinitas · 3 years, 6 months ago
- 655e8c6 Make data_layout an attribute of the Scale function by Michele Di Giorgio · 3 years, 6 months ago
- c9c8905 Make CL Elementwise Unary kernels and functions state-less by Michele Di Giorgio · 3 years, 6 months ago
- c727d52 Add broadcasting support in NEPixelWiseMultiplicationKernel for FP16 by Pablo Marquez Tello · 3 years, 6 months ago
- ef28340 Force early exit of threads when output_height=1 and channels=1 in assembly pooling by Michele Di Giorgio · 3 years, 6 months ago
- 237be03 Make Gpu Elementwise Operations kernels and functions state-less by Michele Di Giorgio · 3 years, 6 months ago
- 26cf798 Fix NEON/CL CustomConvolution failures by Manuel Bottini · 3 years, 6 months ago
- 4cfab18 Make CLArithmeticSubtraction kernel and function state-less by Michele Di Giorgio · 3 years, 6 months ago