diff options
| author | Rob Hughes <robert.hughes@arm.com> | 2019-09-24 16:59:56 +0100 |
|---|---|---|
| committer | Áron Virginás-Tar <aron.virginas-tar@arm.com> | 2019-09-27 10:17:26 +0000 |
| commit | 3a7d3a70d99cbe22f5e4711d5dbbea2a245da7ed (patch) | |
| tree | 445967fa35016374657b5c7e38b2715773a25e02 /Android.mk | |
| parent | 83239f995e7b86062450794b85bfe4c4c387fda0 (diff) | |
| download | armnn-3a7d3a70d99cbe22f5e4711d5dbbea2a245da7ed.tar.gz | |
NNXSW-1826 Add an optimization step which combines Permute and BatchToSpace into DepthToSpace
This is only possible in some limited cases, but removes an extra
layer from the graph and so should improve performance in all cases.
Change-Id: I7b3e6ba5dacb4fdb816ad270edaecda1436ab4cf
Signed-off-by: Rob Hughes <robert.hughes@arm.com>
Diffstat (limited to 'Android.mk')
| -rw-r--r-- | Android.mk | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/Android.mk b/Android.mk index 3e665a039f..4c3789c9e0 100644 --- a/Android.mk +++ b/Android.mk @@ -91,6 +91,7 @@ LOCAL_SRC_FILES := \ src/armnn/NetworkUtils.cpp \ src/armnn/Observable.cpp \ src/armnn/Optimizer.cpp \ + src/armnn/optimizations/PermuteAndBatchToSpaceAsDepthToSpace.cpp \ src/armnn/ProfilingEvent.cpp \ src/armnn/Profiling.cpp \ src/armnn/Runtime.cpp \ |