IntenseWebs/openvino
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[CPU][ARM] Enable FP16 precision for CumSum node (#19176)

Browse Source
This commit is contained in:
Egor Duplenskii 2023-08-14 17:22:43 +02:00 committed by GitHub
parent 600c2d8283
commit 680333b2db
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
src/plugins/intel_cpu/src/nodes/cum_sum.cpp
View File

@ -12,6 +12,7 @@
#include <ie_ngraph_utils.hpp>
#include "cum_sum.h"
#include "utils/bfloat16.hpp"
#include "openvino/core/type/float16.hpp"
using namespace InferenceEngine;
@ -72,7 +73,10 @@ void CumSum::initSupportedPrimitiveDescriptors() {
return;
dataPrecision = getOriginalInputPrecisionAtPort(CUM_SUM_DATA);
if (!one_of(dataPrecision, Precision::I8, Precision::U8, Precision::I16, Precision::BF16, Precision::I32, Precision::FP32, Precision::I64, Precision::U64))
if (!one_of(dataPrecision,
Precision::I8, Precision::U8,
Precision::I16, Precision::I32, Precision::I64, Precision::U64,
Precision::BF16, Precision::FP16, Precision::FP32))
IE_THROW() << errorPrefix << " has unsupported 'data' input precision: " << dataPrecision.name();
if (inputShapes.size() == numOfInputs) {
@ -101,6 +105,7 @@ void CumSum::execute(dnnl::stream strm) {
OV_CASE(Precision::U8, uint8_t),
OV_CASE(Precision::I16, int16_t),
OV_CASE(Precision::BF16, bfloat16_t),
OV_CASE(Precision::FP16, ov::float16),
OV_CASE(Precision::I32, int32_t),
OV_CASE(Precision::FP32, float),
OV_CASE(Precision::I64, int64_t),