1
I Use This!
Activity Not Available

Commits : Listings

Analyzed 12 months ago. based on code collected 12 months ago.
Nov 24, 2024 — Nov 24, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Fix CLDepthwiseConvolutionLayer 3x3 QASYMM8 More... over 5 years ago
Fix tolerance for NEON RNNLayer More... over 5 years ago
Fix tolerance for NEON RNNLayer More... almost 6 years ago
Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part2 More... almost 6 years ago
Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part1 More... almost 6 years ago
Fix unused variable compilation error More... almost 6 years ago
Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part4 More... almost 6 years ago
Integrate MLGO into CLGEMM and CLGEMMLowpMatrixMultiplyCore: Part3 More... almost 6 years ago
Fix segfault in fsrcnn.tflite in GpuAcc More... almost 6 years ago
Make NEON Pooling kernels and functions state-less More... almost 6 years ago
Fix CL Compiler frontend failure for ArgMinMax More... almost 6 years ago
Add mlgo to graph examples More... almost 6 years ago
Add dynamic tensor support to CpuElementwise More... almost 6 years ago
Tweak scheduling use of SQDMULH in quantized AVG pooling More... almost 6 years ago
Make memset/copy functions state-less More... almost 6 years ago
Add DIV to Elementwise layer support at graph level More... almost 6 years ago
Account for eventual padding in Pooling assembly kernels More... almost 6 years ago
Update documentation for OpenCL kernel padding removal More... almost 6 years ago
Make CL Pooling kernels and functions state-less More... almost 6 years ago
Fix OpenCL direct convolution More... almost 6 years ago
Provide utlity function to print ITensorInfo More... almost 6 years ago
Regenerate kernels and update A55 versions More... almost 6 years ago
Remove redundant header from NEElementwiseUnaryLayer More... almost 6 years ago
Make data_layout an attribute of the Scale function More... almost 6 years ago
Force early exit of threads when output_height=1 and channels=1 in assembly pooling More... almost 6 years ago
Add WallClockTimer support for bare metal.
mpekatsoula
as Michalis Spyrou
More... almost 6 years ago
Add broadcasting support in NEPixelWiseMultiplicationKernel for FP16 More... almost 6 years ago
Make Softmax kernels on OpenCL stateless More... almost 6 years ago
Fix NEON/CL CustomConvolution failures More... almost 6 years ago
Make CL Elementwise Unary kernels and functions state-less More... almost 6 years ago