diff options
| author | Gunes Bayir <gunes.bayir@arm.com> | 2023-01-29 13:24:24 +0000 |
|---|---|---|
| committer | Gunes Bayir <gunes.bayir@arm.com> | 2023-02-01 09:59:30 +0000 |
| commit | ae72a46e495742863dba44fcf5fdc673c9d2afbc (patch) | |
| tree | 65bab43d0feddaa66b160ac7dc746651dc7c48de /src/BUILD.bazel | |
| parent | ec320d9fc418e2d95a3a38ce87233397535f467d (diff) | |
| download | ComputeLibrary-ae72a46e495742863dba44fcf5fdc673c9d2afbc.tar.gz | |
Add new operator AddMulAdd for Neon™ backend for Float/Quantized types
This is a fused operator that merges Add + Mul + Add [+ Relu-based-Activation] layers and have an intermediate output after the first Add. It's supported for FP16/32/QASYMM8/QASYMM8_SIGNED data types.
The subsequent Add and Mul are intended for scaling and the coefficients only have one dimension (per channel).
The inputs are
- input1 : nD tensor [X, Y, Z, W, ..]
- input2 : nD tensor [X, Y, Z, W, ..]
- add_coef : 1D tensor [X]
- mul_coef : 1D tensor [X]
The outputs are
- out1 : nD tensor (intermediate output) [X, Y, Z, W, ..]
- out2 : nD tensor (final output) [X, Y, Z, W, ..]
The operation can be summarized as follows:
out1 <- input1 + input2
out2 <- Act(out1 * mul_coef + add_coef)
The activation function can be Identity, Relu, Bounded Relu or Lower/Upper Bounded Relu. The intermediate output can be skipped by providing a nullptr.
The reason of providing this operator is to be able to fuse in case of Residual network patterns and save computations by reducing memory back and forward.
Resolves: COMPMID-5463
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Change-Id: I8ef577aa623b036e9a9f655cc088493fd19a6109
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9055
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/BUILD.bazel')
| -rw-r--r-- | src/BUILD.bazel | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/src/BUILD.bazel b/src/BUILD.bazel index 93f7d1acb9..d8b0127179 100644 --- a/src/BUILD.bazel +++ b/src/BUILD.bazel @@ -653,6 +653,7 @@ filegroup( "cpu/CpuTensor.cpp", "cpu/kernels/CpuActivationKernel.cpp", "cpu/kernels/CpuAddKernel.cpp", + "cpu/kernels/CpuAddMulAddKernel.cpp", "cpu/kernels/CpuCastKernel.cpp", "cpu/kernels/CpuCol2ImKernel.cpp", "cpu/kernels/CpuConcatenateBatchKernel.cpp", @@ -710,6 +711,10 @@ filegroup( "cpu/kernels/add/generic/neon/qasymm8.cpp", "cpu/kernels/add/generic/neon/qasymm8_signed.cpp", "cpu/kernels/add/generic/neon/qsymm16.cpp", + "cpu/kernels/addmuladd/generic/neon/fp16.cpp", + "cpu/kernels/addmuladd/generic/neon/fp32.cpp", + "cpu/kernels/addmuladd/generic/neon/qasymm8.cpp", + "cpu/kernels/addmuladd/generic/neon/qasymm8_signed.cpp", "cpu/kernels/boundingboxtransform/generic/neon/fp16.cpp", "cpu/kernels/boundingboxtransform/generic/neon/fp32.cpp", "cpu/kernels/boundingboxtransform/generic/neon/impl.cpp", @@ -810,6 +815,7 @@ filegroup( "cpu/kernels/sub/neon/qsymm16.cpp", "cpu/operators/CpuActivation.cpp", "cpu/operators/CpuAdd.cpp", + "cpu/operators/CpuAddMulAdd.cpp", "cpu/operators/CpuCast.cpp", "cpu/operators/CpuConcatenate.cpp", "cpu/operators/CpuConv2d.cpp", @@ -867,6 +873,7 @@ filegroup( "runtime/NEON/INESimpleFunction.cpp", "runtime/NEON/INESimpleFunctionNoBorder.cpp", "runtime/NEON/functions/NEActivationLayer.cpp", + "runtime/NEON/functions/NEAddMulAdd.cpp", "runtime/NEON/functions/NEArgMinMaxLayer.cpp", "runtime/NEON/functions/NEArithmeticAddition.cpp", "runtime/NEON/functions/NEArithmeticSubtraction.cpp", |