IVGCVSW-4633 Add conversion of BF16 support to Neon

 * Add NeonConvertBf16ToFp32Workload
 * Add NeonConvertFp32ToBf16Workload
 * Add BFloat16 type support to NeonConstantWorkload and NeonTensorHandle
 * Add ConvertBf16ToFp32Weight when ConvertBf16ToFp32Layer is added
 * Unit tests

Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: Id5b44a203add5e0c98c1ca4e2162115741b56644

1 files changed, 6 insertions, 0 deletions

diff --git a/src/backends/neon/workloads/NeonConstantWorkload.cpp b/src/backends/neon/workloads/NeonConstantWorkload.cpp
index 08f0390e34..83a2692b6e 100644
--- a/src/backends/neon/workloads/NeonConstantWorkload.cpp
+++ b/src/backends/neon/workloads/NeonConstantWorkload.cpp
@@ -6,6 +6,7 @@
 #include "NeonConstantWorkload.hpp"
 
 #include <arm_compute/core/Types.h>
+#include <BFloat16.hpp>
 #include <Half.hpp>
 #include <aclCommon/ArmComputeTensorUtils.hpp>
 #include <neon/NeonTensorHandle.hpp>
@@ -46,6 +47,11 @@ void NeonConstantWorkload::Execute() const
 
         switch (computeDataType)
         {
+            case arm_compute::DataType::BFLOAT16:
+            {
+                CopyArmComputeITensorData(data.m_LayerOutput->GetConstTensor<BFloat16>(), output);
+                break;
+            }
             case arm_compute::DataType::F16:
             {
                 CopyArmComputeITensorData(data.m_LayerOutput->GetConstTensor<Half>(), output);