diff options
| author | Mike Kelly <mike.kelly@arm.com> | 2020-11-12 10:58:48 +0000 |
|---|---|---|
| committer | Jim Flynn <jim.flynn@arm.com> | 2020-11-13 14:25:30 +0000 |
| commit | 07810fc2fcdd34db74222d90cc73ef12a88e7b78 (patch) | |
| tree | 8becef8453674822d079815b06ae37310b97d2cf /src/backends/backendsCommon/IBackendInternal.cpp | |
| parent | 8502adeafbbb1db0acefa62560d93453e38dcadb (diff) | |
| download | armnn-07810fc2fcdd34db74222d90cc73ef12a88e7b78.tar.gz | |
IVGCVSW-5328-5329 Fuse Activation
* Added Fused Activation Optimization to both CL and Neon backends.
* Added Fused Activation support to all the CL and Neon workloads
that support it.
* Changed ProfilingTest network to be a Convolution layer
followed by an Abs layer rather than an Activation layer.
* Added IBackendInternal::OptimizeSubgraphView function that can accept a
ModelOptions.
* Network will now call OptimizeSubgraphView passing in the ModelOptions.
Signed-off-by: Keith Davis <keith.davis@arm.com>
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Ib536ac3cbafc7d9b35c139ad9a65b7735262cd9d
Diffstat (limited to 'src/backends/backendsCommon/IBackendInternal.cpp')
| -rw-r--r-- | src/backends/backendsCommon/IBackendInternal.cpp | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/src/backends/backendsCommon/IBackendInternal.cpp b/src/backends/backendsCommon/IBackendInternal.cpp index 81fc515b98..b08dff84ed 100644 --- a/src/backends/backendsCommon/IBackendInternal.cpp +++ b/src/backends/backendsCommon/IBackendInternal.cpp @@ -3,6 +3,7 @@ // SPDX-License-Identifier: MIT // +#include <armnn/BackendOptions.hpp> #include <armnn/backends/IBackendInternal.hpp> namespace armnn @@ -135,6 +136,12 @@ OptimizationViews IBackendInternal::OptimizeSubgraphView(const SubgraphView& sub return result; } +OptimizationViews IBackendInternal::OptimizeSubgraphView(const SubgraphView& subgraph, + const ModelOptions& /*modelOptions*/) const +{ + return OptimizeSubgraphView(subgraph); +} + bool IBackendInternal::SupportsTensorAllocatorAPI() const { return !GetHandleFactoryPreferences().empty(); |