diff options
| author | Francis Murtagh <francis.murtagh@arm.com> | 2020-03-10 13:51:45 +0000 |
|---|---|---|
| committer | Jim Flynn <jim.flynn@arm.com> | 2020-03-10 14:02:00 +0000 |
| commit | ddb1d06dbcb5dc4a89a237ac1176279669817f46 (patch) | |
| tree | e2d1104bea631f8f366f529201041e701d09b998 /src/backends/backendsCommon/WorkloadData.cpp | |
| parent | 6445cfff7519effd1df04eac88ae17d6e4e6693b (diff) | |
| download | armnn-ddb1d06dbcb5dc4a89a237ac1176279669817f46.tar.gz | |
MLCE-159 Add QAsymmS8 to ArmnnQuantizer
* Allow per layer quantization from Fp32 to Int8 (QAsymmS8) like TfLite
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: I5bbf770aa29d81af3568c15b47d2b2c18e55bb28
Diffstat (limited to 'src/backends/backendsCommon/WorkloadData.cpp')
| -rw-r--r-- | src/backends/backendsCommon/WorkloadData.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/backends/backendsCommon/WorkloadData.cpp b/src/backends/backendsCommon/WorkloadData.cpp index dbd1158380..bb0c21ffba 100644 --- a/src/backends/backendsCommon/WorkloadData.cpp +++ b/src/backends/backendsCommon/WorkloadData.cpp @@ -994,6 +994,7 @@ void FullyConnectedQueueDescriptor::Validate(const WorkloadInfo& workloadInfo) c { DataType::Float32, DataType::Float16, + DataType::QAsymmS8, DataType::QAsymmU8, DataType::QSymmS16 }; @@ -1183,8 +1184,8 @@ void Convolution2dQueueDescriptor::Validate(const WorkloadInfo& workloadInfo) co std::vector<DataType> supportedTypes = { DataType::Float32, - DataType::QAsymmU8, DataType::QAsymmS8, + DataType::QAsymmU8, DataType::QSymmS16, DataType::QSymmS8, DataType::Float16 |