- 1af5416 Add experimental PostOp interface to ClGemmMatrixMultiplyReshapedKernel Part 1 by SiCongLi · 2 years, 11 months ago
- 69df64f Improve conv3d validation by Freddie Liardet · 2 years, 10 months ago
- f727ef4 Add uint8/int8 support to cpu conv3d by Freddie Liardet · 2 years, 10 months ago
- 2277a0d Fix enforcing incorrect ISA feature for given CPU models by Georgios Pinitas · 2 years, 11 months ago
- 51847d5 Implement CLDirectConv3DKernel - uint8/int8 by Giorgio Arena · 2 years, 10 months ago
- 5dda217 DirectConv3d support refine by Sheri Zhang · 2 years, 11 months ago
- c9cecc0 Remove legacy GeMM kernels on OpenCL by Gian Marco Iodice · 2 years, 11 months ago
- 3611852 Fix precision issue in ChannelShuffleKernel by Pablo Marquez Tello · 2 years, 10 months ago
- 487d390 Add user provided JSON operator list build by Freddie Liardet · 3 years ago
- 4cdd6b8 Implement Minimum Workload Size (MWS) in all CPPKernels used by small networks by Dana Zlotnik · 2 years, 11 months ago
- 6d9c982 Conv3d support by Sheri Zhang · 3 years ago
- 273c28c Fix CLConv3D filelist and comments by Giorgio Arena · 2 years, 11 months ago
- 945ae9e Implement CLDirectConv3D f32/f16 by Giorgio Arena · 2 years, 11 months ago
- de23fc3 Removed unsupported NHWC from Pooling::validate by Pablo Tello · 2 years, 11 months ago
- b060806 Improve performance of Softmax uint8 on GPU by Adnan AlSinan · 2 years, 11 months ago
- ded3663 Remove padding in cpuPool2d NCHW by Freddie Liardet · 3 years ago
- 0d11b70 Fix precision issue in ChannelShuffleKernel by Pablo Tello · 2 years, 11 months ago
- c9fe9fc Add support for 5D data layout indexing by Giorgio Arena · 2 years, 11 months ago
- b6af482 Per-operator build dependencies by Georgios Pinitas · 3 years ago
- 58e9e06 Fix the multipliers calculation by Pablo Marquez Tello · 2 years, 11 months ago
- 1024d9f Add missing ISA selection heuristics by Georgios Pinitas · 2 years, 11 months ago
- 8b8405a Optimize CpuScale NHWC F32/F16 by Gian Marco Iodice · 2 years, 11 months ago
- 8229e5a Fix of linking errors to find CPPSplit symbols when compiling android nn driver by ramelg01 · 2 years, 11 months ago
- 41d886c Fixed the compiler warning -Werror=type-limits by Pablo Marquez Tello · 2 years, 11 months ago
- 4a6d9e8 Provide logging for configure functions in all CPP functions by ramelg01 · 2 years, 11 months ago
- 6d89157 Provide logging for configure functions in all CL functions by ramelg01 · 2 years, 11 months ago
- 63e0beb Add support for non-constant weights and biases in CpuFullyConnected by Giorgio Arena · 3 years ago
- 2e53f17 Provide logging for configure functions in all gpu operators by ramelg01 · 3 years ago
- 0bb6230 Fix inefficient store in gemmlowp_mm_reshaped_only_rhs_t by Gian Marco Iodice · 3 years ago
- 79cb945 Update OpenCL header file to version 2020.12.18 by Sheri Zhang · 3 years ago
- cbbb038 Provide logging for configure functions in all NEON functions by ramelg01 · 3 years ago
- 3ae3d88 Provide logging for configure functions in all cpu operators by ramelg01 · 3 years ago
- 9ac7b99 Revert "Add support for non-constant weights and biases in CpuFullyConnected" by Pablo Marquez Tello · 3 years ago
- cc171f9 Provide tests for fusing pad layer in graph API by Gunes Bayir · 3 years ago
- e4563a0 Adds Conv3d reference implementation support. by Adnan AlSinan · 3 years ago
- e42a87f Update cpu heuristics for GEMM by Gunes Bayir · 3 years ago
- 7bc1a77 Optimize ClScaleKernel on NHWC (f32/f16/int8) by Gian Marco Iodice · 3 years ago
- 951d520 Remove padding from ClGemmMatrixMultiplyReshapedOnlyRhsKernel by Giorgio Arena · 3 years ago
- 2f9ae16 Avoid checking on biases' constantness if nullptr by Giorgio Arena · 3 years ago
- 88bd35d Fix vload_partial macros on OpenCL by Giorgio Arena · 3 years ago
- bde2f35 Remove padding from ClGemmMatrixMultiplyReshapedKernel by Giorgio Arena · 3 years ago
- 814bddf Fuse pad layer with subsequent convolution layer by Gunes Bayir · 3 years ago
- aed63ee Add support for non-constant weights and biases in CpuFullyConnected by Michele Di Giorgio · 3 years, 1 month ago
- d216f57 Update cpu depthwise kernels by Freddie Liardet · 3 years, 1 month ago
- 4e53c5a Revert "Remove padding from ClGemmMatrixMultiplyReshapedKernel" by Pablo Marquez Tello · 3 years ago
- 8afe6c7 OpenCL job chaining support by Giorgio Arena · 3 years ago
- 8fce496 Remove padding from ClPool2dKernel NCHW by Giorgio Arena · 3 years ago
- aed61f2 Fix CLNormalizationLayer NCHW border calculation by SiCongLi · 3 years ago
- 50335fd Remove padding from ClGemmMatrixMultiplyReshapedKernel by Michele Di Giorgio · 3 years ago
- e920d6a Printing operators parameters, currently for CpuAdd operator only. by Ramy Elgammal · 3 years ago
- c89998f Fix strict overflow warnings by Michalis Spyrou · 3 years ago
- 7891a73 Move CPU/GPU files from Core/Runtime to the respective backend folders by Georgios Pinitas · 3 years ago
- a46c9c9 Sync backend only if allocated rather than supported by Giorgio Arena · 3 years ago
- 9c67d38 Remove map/unmap overhead for input/output accessor when using DummyAccessor by Giorgio Arena · 3 years ago
- 87a74ef Re-use auxiliary memory withing CpuWinogradConv2d operators by Georgios Pinitas · 3 years ago
- 511771f Remove padding from ClScaleKernel by Giorgio Arena · 3 years ago
- 1988463 Rename [Cl|Cpu]GemmConvolution to [Cl|Gpu]GemmConv2d by Georgios Pinitas · 3 years ago
- 73df931 Address comments on avoiding releasing weights if used by multiple functions by Giorgio Arena · 3 years ago
- df5bcb6 Enable fast_math on CpuGemmConvolution by Georgios Pinitas · 3 years ago
- ebbd529 Update the heuristic to call direct convolution in clConv2d by Gian Marco Iodice · 3 years ago
- f5d51f3 Retain weights in ClGemm when reconfiguring the operator with retention by Georgios Pinitas · 3 years ago
- 45fbf9f Avoid releasing weights if they are used by multiple functions by Georgios Pinitas · 3 years ago
- d4a5bc5 Ensure correct transformed matrices are used in CpuGemmConvolution by Georgios Pinitas · 3 years, 1 month ago
- fa1db17 Ensure that correct transformed matrices are used in CpuFullyConnected by Georgios Pinitas · 3 years, 1 month ago
- d761a3e Fix performance regression due to clFinish() by Gian Marco Iodice · 3 years, 1 month ago
- 288d3cb Fix compiler error in CLActivationLayer by Pablo Marquez Tello · 3 years, 1 month ago
- acb4d93 Fix compiler error in GCC 7.4 + Ubuntu 16 by Pablo Marquez Tello · 3 years, 1 month ago
- 3f74389 Remove 21.08 deprecated code by Freddie Liardet · 3 years, 1 month ago
- 0f6ca4b Report error for unsupported non-constant weights in CpuFullyConnected by Michele Di Giorgio · 3 years, 1 month ago
- 6269f87 Fix depthwise convolution assembly kernels by Freddie Liardet · 3 years, 1 month ago
- 6634194 Avoid over-allocation of temporary buffers within CpuWinogradConv2d by Georgios Pinitas · 3 years, 1 month ago
- 4164814 Implement Operator API by Georgios Pinitas · 3 years, 1 month ago
- 6e90c12 Add missing limits include by Freddie Liardet · 3 years, 1 month ago
- ba2dd82 Benchmark and set default LWS for GEMM, Direct convolution and Winograd by Giorgio Arena · 3 years, 1 month ago
- 06d1efd Port CLConvolutionLayer by Sheri Zhang · 3 years, 1 month ago
- 529b5a2 Port ClFullyConnected to new API by Georgios Pinitas · 3 years, 1 month ago
- a0efe69 Reintroduce implementation of NEConvolutionLayer::get_convolution_method by Michele Di Giorgio · 3 years, 1 month ago
- 9a1e28b Compilation issue: neon=1 armv8.2 on Android with NDKr18beta1 by Gian Marco Iodice · 3 years, 1 month ago
- a76e403 Fix A55 performance constant for fp16 hybrid gemm kernel by Georgios Pinitas · 3 years, 1 month ago
- b55f8e8 Port NEConvolutionLayer by Michalis Spyrou · 3 years, 1 month ago
- 8e2f64f Create custom flags for enabling fp16 support by Georgios Pinitas · 3 years, 1 month ago
- a86433a Reduce binary footprint of CpuConvertFullyConnectedWeightsKernel by Michele Di Giorgio · 3 years, 1 month ago
- 120c87b Fix bare metal build issues by Freddie Liardet · 3 years, 1 month ago
- bb5d42c Fix cpu GEMM fp16 issue by Freddie Liardet · 3 years, 1 month ago
- f81f51c Reorganize the kernels into nhwc, nchw and common folders by Adnan AlSinan · 3 years, 1 month ago
- 97426a7 Remove generated kernels that overlap hand-written ones by Georgios Pinitas · 3 years, 1 month ago
- 9805583 Fix memory lifetime issue by Georgios Pinitas · 3 years, 1 month ago
- d87aded Port CLGEMMConvolutionLayer by Manuel Bottini · 3 years, 1 month ago
- 4718706 Dispatch Conv2d using the Direct method when necessary by Georgios Pinitas · 3 years, 1 month ago
- 387f80e Update GEMM assembly performance parameters by Georgios Pinitas · 3 years, 1 month ago
- 5e61cb0 Add missing limits include by Freddie Liardet · 3 years, 1 month ago
- e92c23e Fix allocation of prepare tensor on ClWinogradConv2d by Georgios Pinitas · 3 years, 1 month ago
- 7075fe2 Reorganize the kernels into nhwc, nchw and common folders by Adnan AlSinan · 3 years, 2 months ago
- 22f5ed5 Avoid allocation of auxiliary memory in CpuGemmConvolution by Georgios Pinitas · 3 years, 1 month ago
- 8bd2556 Fix vector_length identification mechanism for SVE by Georgios Pinitas · 3 years, 1 month ago
- d9cdf14 Port NEFullyConnectedLayer to memory injecting interface by Michele Di Giorgio · 3 years, 2 months ago
- a8297fb Pass fast math flag for correct GEMM3D validation support by Georgios Pinitas · 3 years, 1 month ago
- e7012a6 Fix bare metal build error by Freddie Liardet · 3 years, 1 month ago
- 69a9ac4 Expose fast_math mode for GEMM through BFloat16 by Georgios Pinitas · 3 years, 1 month ago
- 1fa27ad Inject temporary tensors to pack in they don't exist in CpuSoftmax by Georgios Pinitas · 3 years, 1 month ago