commit 71ecf396bb08e302dc06b2c7ed44001894d3a2ea
author Vidhya Sudhan Loganathan <vidhyasudhan.loganathan@arm.com> Fri Aug 31 16:10:16 2018 +0100
committer Anthony Barbier <anthony.barbier@arm.com> Fri Nov 02 16:54:54 2018 +0000
tree 41a0143c3acd77c9c995c7d97ade143e49719392
parent 553b999ccc4233b163377e0a55e2377614899a3e

COMPMID-1266 : support for FP16 in CLWinogradConvolutionLayer

Added support for FP16 in CLWinogradConvolutionLayer: 5x5 kernels and 3x3 kernels(COMPMID-937)

Change-Id: I0f394cbdc978dd04176416e9f612aca3986b09e6
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/145537
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>

tests/validation/Helpers.cpp

diff --git a/tests/validation/Helpers.cpp b/tests/validation/Helpers.cpp
index 1044e32..fd034b6 100644
--- a/tests/validation/Helpers.cpp
+++ b/tests/validation/Helpers.cpp
@@ -169,7 +169,8 @@
     return dst;
 }
 
-void matrix_multiply(const SimpleTensor<float> &a, const SimpleTensor<float> &b, SimpleTensor<float> &out)
+template <typename T>
+void matrix_multiply(const SimpleTensor<T> &a, const SimpleTensor<T> &b, SimpleTensor<T> &out)
 {
     ARM_COMPUTE_ERROR_ON(a.shape()[0] != b.shape()[1]);
     ARM_COMPUTE_ERROR_ON(a.shape()[1] != out.shape()[1]);
@@ -194,7 +195,8 @@
     }
 }
 
-void transpose_matrix(const SimpleTensor<float> &in, SimpleTensor<float> &out)
+template <typename T>
+void transpose_matrix(const SimpleTensor<T> &in, SimpleTensor<T> &out)
 {
     ARM_COMPUTE_ERROR_ON((in.shape()[0] != out.shape()[1]) || (in.shape()[1] != out.shape()[0]));
 
@@ -301,7 +303,14 @@
 }
 
 template void get_tile(const SimpleTensor<float> &in, SimpleTensor<float> &roi, const Coordinates &coord);
+template void get_tile(const SimpleTensor<half> &in, SimpleTensor<half> &roi, const Coordinates &coord);
 template void zeros(SimpleTensor<float> &in, const Coordinates &anchor, const TensorShape &shape);
+template void zeros(SimpleTensor<half> &in, const Coordinates &anchor, const TensorShape &shape);
+template void transpose_matrix(const SimpleTensor<float> &in, SimpleTensor<float> &out);
+template void transpose_matrix(const SimpleTensor<half> &in, SimpleTensor<half> &out);
+template void matrix_multiply(const SimpleTensor<float> &a, const SimpleTensor<float> &b, SimpleTensor<float> &out);
+template void matrix_multiply(const SimpleTensor<half> &a, const SimpleTensor<half> &b, SimpleTensor<half> &out);
+
 } // namespace validation
 } // namespace test
 } // namespace arm_compute