index
:
ComputeLibrary.git
branches/arm_compute_19_02
branches/arm_compute_19_05
branches/arm_compute_19_08
branches/arm_compute_19_11
branches/arm_compute_20_02
branches/arm_compute_20_05
branches/arm_compute_20_08
branches/arm_compute_20_11
branches/arm_compute_21_02
branches/arm_compute_21_05
branches/arm_compute_21_08
branches/arm_compute_21_11
branches/arm_compute_22_02
branches/arm_compute_22_05
branches/arm_compute_22_08
branches/arm_compute_22_11
branches/arm_compute_23_02
branches/arm_compute_23_02_1
branches/arm_compute_23_05
branches/arm_compute_23_05_1
branches/arm_compute_23_08
branches/arm_compute_23_11
branches/arm_compute_24_01
branches/arm_compute_24_02
branches/arm_compute_24_02_1
branches/arm_compute_24_04
branches/arm_compute_24_05
branches/arm_compute_24_06
branches/arm_compute_24_07
branches/arm_compute_24_08
dev/21_02_int8_optim
dev/21_05_int8_optim
main
master
release_candidate
about
summary
refs
log
tree
commit
diff
log msg
author
committer
range
path:
root
/
src
Age
Commit message (
Expand
)
Author
2021-08-25
Sync backend only if allocated rather than supported
Giorgio Arena
2021-08-24
Remove map/unmap overhead for input/output accessor when using DummyAccessor
Giorgio Arena
2021-08-24
Re-use auxiliary memory withing CpuWinogradConv2d operators
Georgios Pinitas
2021-08-23
Remove padding from ClScaleKernel
Giorgio Arena
2021-08-20
Rename [Cl|Cpu]GemmConvolution to [Cl|Gpu]GemmConv2d
Georgios Pinitas
2021-08-19
Address comments on avoiding releasing weights if used by multiple functions
Giorgio Arena
2021-08-18
Enable fast_math on CpuGemmConvolution
Georgios Pinitas
2021-08-18
Update the heuristic to call direct convolution in clConv2d
Gian Marco Iodice
2021-08-18
Retain weights in ClGemm when reconfiguring the operator with retention
Georgios Pinitas
2021-08-13
Avoid releasing weights if they are used by multiple functions
Georgios Pinitas
2021-08-13
Ensure correct transformed matrices are used in CpuGemmConvolution
Georgios Pinitas
2021-08-12
Ensure that correct transformed matrices are used in CpuFullyConnected
Georgios Pinitas
2021-08-11
Fix performance regression due to clFinish()
Gian Marco Iodice
2021-08-10
Fix compiler error in CLActivationLayer
Pablo Marquez Tello
2021-08-06
Fix compiler error in GCC 7.4 + Ubuntu 16
Pablo Marquez Tello
2021-08-04
Remove 21.08 deprecated code
Freddie Liardet
2021-08-04
Report error for unsupported non-constant weights in CpuFullyConnected
Michele Di Giorgio
2021-08-04
Fix depthwise convolution assembly kernels
Freddie Liardet
2021-08-04
Avoid over-allocation of temporary buffers within CpuWinogradConv2d
Georgios Pinitas
2021-08-04
Implement Operator API
Georgios Pinitas
2021-08-02
Add missing limits include
Freddie Liardet
2021-08-02
Benchmark and set default LWS for GEMM, Direct convolution and Winograd
Giorgio Arena
2021-08-02
Port CLConvolutionLayer
Sheri Zhang
2021-07-30
Port ClFullyConnected to new API
Georgios Pinitas
2021-07-30
Reintroduce implementation of NEConvolutionLayer::get_convolution_method
Michele Di Giorgio
2021-07-30
Compilation issue: neon=1 armv8.2 on Android with NDKr18beta1
Gian Marco Iodice
2021-07-29
Fix A55 performance constant for fp16 hybrid gemm kernel
Georgios Pinitas
2021-07-29
Port NEConvolutionLayer
Michalis Spyrou
2021-07-28
Create custom flags for enabling fp16 support
Georgios Pinitas
2021-07-28
Reduce binary footprint of CpuConvertFullyConnectedWeightsKernel
Michele Di Giorgio
2021-07-28
Fix bare metal build issues
Freddie Liardet
2021-07-28
Fix cpu GEMM fp16 issue
Freddie Liardet
2021-07-28
Reorganize the kernels into nhwc, nchw and common folders
Adnan AlSinan
2021-07-28
Remove generated kernels that overlap hand-written ones
Georgios Pinitas
2021-07-27
Fix memory lifetime issue
Georgios Pinitas
2021-07-27
Port CLGEMMConvolutionLayer
Manuel Bottini
2021-07-27
Dispatch Conv2d using the Direct method when necessary
Georgios Pinitas
2021-07-27
Update GEMM assembly performance parameters
Georgios Pinitas
2021-07-26
Add missing limits include
Freddie Liardet
2021-07-26
Fix allocation of prepare tensor on ClWinogradConv2d
Georgios Pinitas
2021-07-25
Reorganize the kernels into nhwc, nchw and common folders
Adnan AlSinan
2021-07-23
Avoid allocation of auxiliary memory in CpuGemmConvolution
Georgios Pinitas
2021-07-23
Fix vector_length identification mechanism for SVE
Georgios Pinitas
2021-07-23
Port NEFullyConnectedLayer to memory injecting interface
Michele Di Giorgio
2021-07-23
Pass fast math flag for correct GEMM3D validation support
Georgios Pinitas
2021-07-23
Fix bare metal build error
Freddie Liardet
2021-07-22
Expose fast_math mode for GEMM through BFloat16
Georgios Pinitas
2021-07-22
Inject temporary tensors to pack in they don't exist in CpuSoftmax
Georgios Pinitas
2021-07-22
Port ClGemmLowp to new API
Georgios Pinitas
2021-07-22
Fix oclgrind int overflow warning
Freddie Liardet
[next]