Skip to content

N
ngraph
Commit 310fcf07 authored by mbencer's avatar mbencer
Browse files
Options

Merge remote-tracking branch 'origin/master' into mbencer/BuilderSplitV1

parents b4b352d5 6c0cf85a
Hide whitespace changes
Inline Side-by-side
Showing with 826 additions and 291 deletions
cmake/external_mkldnn_v1.cmake
View file @ 310fcf07
......@@ -18,12 +18,12 @@ include(ExternalProject)
# Includes blas 3.8.0 in mkldnn
set(NGRAPH_MKLDNN_SHORT_VERSION 1)
set(NGRAPH_MKLDNN_FULL_VERSION 1.1.1.0)
set(NGRAPH_MKLDNN_FULL_VERSION 1.2.0.0)
set(NGRAPH_MKLDNN_MKLML_ASSET_VERSION "v0.21")
set(NGRAPH_MKLDNN_VERSION "v1.1.1")
set(NGRAPH_MKLDNN_VERSION "v1.2")
set(NGRAPH_MKLDNN_MKLML_VERSION "2019.0.5.20190502")
set(NGRAPH_MKLDNN_MKLML_WIN32_VERSION "2020.0.20190813")
set(NGRAPH_MKLDNN_GIT_TAG "v1.1.1")
set(NGRAPH_MKLDNN_GIT_TAG "v1.2")
#------------------------------------------------------------------------------
# Fetch and install MKL-DNN
......
cmake/external_mlir.cmake
View file @ 310fcf07
......@@ -19,7 +19,7 @@ include(ExternalProject)
set(MLIR_LLVM_REPO_URL https://github.com/llvm/llvm-project.git)
# Change these commit IDs to move to latest stable versions
set(MLIR_LLVM_COMMIT_ID 96400ae)
set(MLIR_LLVM_COMMIT_ID 376c6853)
# MLIR environment variables. Some of them are used by LIT tool.
......
src/contrib/mlir/backend/analysis/memory_analysis.cpp
View file @ 310fcf07
......@@ -26,7 +26,6 @@
#include <llvm/ADT/DenseSet.h>
#include <map>
#include <mlir/EDSC/Builders.h>
#include <mlir/EDSC/Helpers.h>
#include <mlir/EDSC/Intrinsics.h>
#include <mlir/IR/AffineExpr.h>
#include <mlir/IR/IntegerSet.h>
......
src/contrib/mlir/backend/cpu/cpu_backend.cpp
View file @ 310fcf07
......@@ -194,7 +194,8 @@ void MLIRCPUBackend::lowerNgDialect()
void MLIRCPUBackend::lowerStandardDialect()
{
mlir::PassManager pm(&m_context);
pm.addPass(mlir::createLowerToLLVMPass());
pm.addPass(mlir::createLowerToLLVMPass(
/*useAlloca=*/false, /*useBarePtrCallConv=*/false, /*emitCWrappers=*/true));
// Apply any generic pass manager command line options.
mlir::applyPassManagerCLOptions(pm);
......
src/contrib/mlir/backend/pass/affine_lowerer.cpp
View file @ 310fcf07
......@@ -28,9 +28,9 @@
#include <llvm/ADT/DenseSet.h>
#include <llvm/Support/Debug.h>
#include <mlir/EDSC/Builders.h>
#include <mlir/EDSC/Helpers.h>
#include <mlir/EDSC/Intrinsics.h>
#include <mlir/Dialect/AffineOps/EDSC/Builders.h>
#include <mlir/Dialect/AffineOps/EDSC/Intrinsics.h>
#include <mlir/Dialect/StandardOps/EDSC/Intrinsics.h>
#include <mlir/IR/AffineExpr.h>
#include <mlir/IR/Function.h>
#include <mlir/IR/IntegerSet.h>
......@@ -51,11 +51,10 @@ namespace
{
using namespace mlir;
using namespace mlir::edsc;
using namespace mlir::edsc::intrinsics;
using namespace mlir::edsc::op;
using namespace ngraph::runtime;
using namespace ngraph::runtime::ngmlir;
// Index notation to generate standard (i.e., non-affine) loads and stores.
using StdIndexedValue = TemplatedIndexedValue<intrinsics::std_load, intrinsics::std_store>;
class DialectLoweringPass;
......@@ -215,9 +214,37 @@ namespace
NGraphTypeConverter()
: TypeConverter()
{
}
// TODO(dcaballe): split this into independent conversion patterns when there is a
// way to check if a type is valid in Std dialect.
addConversion([this](Type type) -> Type {
if (auto tensorType = type.dyn_cast<NGTensorType>())
{
// Convert NGTensorType to Std MemRefType directly instead of going to Std
// TensorType. This may change in the future.
return MemRefType::get(tensorType.getShape(),
convertType(tensorType.getElementType()),
{/* no map used */},
0);
}
if (auto floatType = type.dyn_cast<NGFloatType>())
{
// Float types are already std type.
return floatType;
}
if (auto intType = type.dyn_cast<NGIntegerType>())
{
return mlir::IntegerType::get(intType.getWidth(), intType.getContext());
}
if (auto boolType = type.dyn_cast<NGBoolType>())
{
return mlir::IntegerType::get(1 /* width */, boolType.getContext());
}
Type convertType(Type t) override;
// Do not assert/NGRAPH_CHECK here. Type convertion infra expects `convertType` to
// return the input type if the type is not supported.
return type;
});
}
};
/// Dialect Lowering Pass to affine ops
......@@ -317,7 +344,8 @@ namespace
// TODO: Encode no alias attribute as part of the function signature conversion or as a
// separate rewrite pattern. Retrieve new function after signature conversion.
insertNoAliasArgAttrs();
// TODO: To be enabled in follow-up commit.
// insertNoAliasArgAttrs();
}
opAttrsVec = m_attrsVec;
......@@ -492,22 +520,22 @@ namespace
/// Add llvm.noalias attribute to all the memref function arguments. We know that this is safe
/// by nGraph op semantics.
void DialectLoweringPass::insertNoAliasArgAttrs()
{
FuncOp func = getModule().lookupSymbol<mlir::FuncOp>(funcName);
NGRAPH_CHECK(func, "FuncOp '" + funcName.str() + "' not found");
unsigned int argIdx = 0;
for (auto arg : func.getArguments())
{
if (arg.getType().isa<MemRefType>())
{
func.setArgAttr(argIdx, "llvm.noalias", BoolAttr::get(true, &getContext()));
}
++argIdx;
}
}
// void DialectLoweringPass::insertNoAliasArgAttrs()
//{
// FuncOp func = getModule().lookupSymbol<mlir::FuncOp>(funcName);
// NGRAPH_CHECK(func, "FuncOp '" + funcName.str() + "' not found");
// unsigned int argIdx = 0;
// for (auto arg : func.getArguments())
// {
// if (arg.getType().isa<MemRefType>())
// {
// func.setArgAttr(argIdx, "llvm.noalias", BoolAttr::get(true, &getContext()));
// }
// ++argIdx;
// }
//}
void DialectLoweringPass::insertDeallocs(PatternRewriter& rewriter)
{
......@@ -543,40 +571,6 @@ namespace
return m_attrsVec.size() - 1;
}
// NGDialect converters
Type NGraphTypeConverter::convertType(Type type)
{
// We may need to refactor this code to a external utility if type conversion is needed
// outside of the lowering context since NGraphTypeConverter is private.
if (auto tensorType = type.dyn_cast<NGTensorType>())
{
// Convert NGTensorType to Std MemRefType directly instead of going to Std TensorType.
// This may change in the future.
return MemRefType::get(tensorType.getShape(),
convertType(tensorType.getElementType()),
{/* no map used */},
0);
}
if (auto floatType = type.dyn_cast<NGFloatType>())
{
// Float types are already std type.
return floatType;
}
if (auto intType = type.dyn_cast<NGIntegerType>())
{
return mlir::IntegerType::get(intType.getWidth(), intType.getContext());
}
if (auto boolType = type.dyn_cast<NGBoolType>())
{
return mlir::IntegerType::get(1 /* width */, boolType.getContext());
}
// Do not assert/NGRAPH_CHECK here. Type convertion infra expects `convertType` to return
// the input type if the type is not supported.
return type;
}
#define REWRITER(OP) \
PatternMatchResult OP##Conversion::matchAndRewrite( \
Operation* op, ArrayRef<Value> operands, ConversionPatternRewriter& rewriter) const
......@@ -680,15 +674,15 @@ namespace
ScopedContext scope(rewriter, loc);
// Views
MemRefView vRes(result), vLHS(lhs);
MemRefBoundsCapture vRes(result), vLHS(lhs);
// Index Values
IndexedValue iRes(result), iLHS(lhs);
AffineIndexedValue iRes(result), iLHS(lhs);
// Bounds Index Handles
auto lbs = vLHS.getLbs();
auto ubs = vLHS.getUbs();
// Loop induction vars
auto ivs = makeIndexHandles(vLHS.rank());
auto pivs = makeHandlePointers(MutableArrayRef<IndexHandle>(ivs));
auto ivs = ValueHandle::makeIndexHandles(vLHS.rank());
auto pivs = makeHandlePointers(ivs);
// Steps
auto steps = vLHS.getSteps();
......@@ -698,7 +692,7 @@ namespace
AffineLoopNestBuilder(pivs, lbs, ubs, steps)([&] {
ValueHandle val = iLHS(ivs);
ValueHandle zero = createZeroConstant(elemTy);
iRes(ivs) = intrinsics::select(val > zero, val, zero);
iRes(ivs) = std_select(val > zero, val, zero);
});
rewriter.replaceOp(op, {result});
......@@ -742,36 +736,37 @@ namespace
// res[n, k] += lhs[n, m] * rhs[m, k]
// TODO (dcab): We currently generate a super naive loop nest. Improve loop nest layout.
MemRefView vRes(result), vLhs(lhs), vRhs(rhs);
MemRefBoundsCapture vRes(result), vLhs(lhs), vRhs(rhs);
NGRAPH_CHECK(vLhs.rank() == 2 && vRhs.rank() == 2 && vRes.rank() == 2,
"Dot operation is only supported for 2D tensors");
// Create induction variables, lower bounds, upper bounds and steps of the loop nest.
// It's important to note that MemRefView priovides lb/ub/step info is "reverse order",
// i.e., fastest varying dimension is the last one, slowest varying dimention is the first
// one.
IndexHandle n, m, k;
// It's important to note that MemRefBoundsCapture priovides lb/ub/step info is "reverse
// order", i.e., fastest varying dimension is the last one, slowest varying dimention is the
// first one.
auto indexType = IndexType::get(rewriter.getContext());
ValueHandle n(indexType), m(indexType), k(indexType);
unsigned nDim = vLhs.fastestVarying() - 1;
unsigned mDim = vRhs.fastestVarying();
unsigned kDim = vRhs.fastestVarying();
IndexHandle nLb(vLhs.lb(nDim)), mLb(vLhs.lb(mDim)), kLb(vRhs.lb(kDim));
IndexHandle nUb(vLhs.ub(nDim)), mUb(vLhs.ub(mDim)), kUb(vRhs.ub(kDim));
ValueHandle nLb(vLhs.lb(nDim)), mLb(vLhs.lb(mDim)), kLb(vRhs.lb(kDim));
ValueHandle nUb(vLhs.ub(nDim)), mUb(vLhs.ub(mDim)), kUb(vRhs.ub(kDim));
int64_t nStep = vLhs.step(nDim), mStep = vLhs.step(mDim), kStep = vRhs.step(kDim);
// Constants and indexed values to be used inside the loop nest.
IndexedValue iRes(result), iLhs(lhs), iRhs(rhs);
AffineIndexedValue iRes(result), iLhs(lhs), iRhs(rhs);
ValueHandle zeroInit(rewriter.create<ConstantOp>(loc, rewriter.getZeroAttr(elemTy)));
{
IndexHandle n, k;
LoopBuilder::makeAffine(&n, nLb, nUb, nStep)([&] {
LoopBuilder::makeAffine(&k, kLb, kUb, kStep)([&] { iRes(n, k) = zeroInit; });
ValueHandle n(indexType), k(indexType);
makeAffineLoopBuilder(&n, nLb, nUb, nStep)([&] {
makeAffineLoopBuilder(&k, kLb, kUb, kStep)([&] { iRes(n, k) = zeroInit; });
});
}
LoopBuilder::makeAffine(&n, nLb, nUb, nStep)([&] {
LoopBuilder::makeAffine(&m, mLb, mUb, mStep)([&] {
LoopBuilder::makeAffine(&k, kLb, kUb, kStep)(
makeAffineLoopBuilder(&n, nLb, nUb, nStep)([&] {
makeAffineLoopBuilder(&m, mLb, mUb, mStep)([&] {
makeAffineLoopBuilder(&k, kLb, kUb, kStep)(
[&] { iRes(n, k) += iLhs(n, m) * iRhs(m, k); });
});
});
......@@ -792,13 +787,13 @@ namespace
NGRAPH_CHECK(result, "Unexpected null result in ConcatOp");
// Create view to write into result.
MemRefView vRes(result);
MemRefBoundsCapture vRes(result);
auto rank = vRes.rank();
// For each operand, generate a separate loop to copy into the target slice of "result".
// We'll keep track of the slice offsets via concatenation_axis_pos.
auto concatenationAxis = concat.concatenation_axis().getSExtValue();
IndexHandle concatenationAxisPos(index_type(0));
Value concatenationAxisPos(std_constant_index(0));
for (auto& operand : operands)
{
......@@ -817,7 +812,7 @@ namespace
// [i_(r-2)][i_(r-1)]
// :=
// operand[i_0][i_1]...[i_(r-2)][i_(r-1)]
MemRefView vOperand(operand);
MemRefBoundsCapture vOperand(operand);
NGRAPH_CHECK(vOperand.rank() == rank, "Unexpected rank mismatch");
llvm::SmallVector<ValueHandle, 5> indexVars;
......@@ -825,9 +820,10 @@ namespace
llvm::SmallVector<ValueHandle, 5> indexVarLbs;
llvm::SmallVector<ValueHandle, 5> indexVarUbs;
llvm::SmallVector<int64_t, 5> indexVarSteps;
auto indexType = IndexType::get(rewriter.getContext());
for (int i = 0; i < rank; i++)
{
indexVars.push_back(IndexHandle());
indexVars.push_back(ValueHandle(indexType));
indexVarPtrs.push_back(&(indexVars.back()));
indexVarLbs.push_back(vOperand.lb(i));
indexVarUbs.push_back(vOperand.ub(i));
......@@ -835,15 +831,15 @@ namespace
}
AffineLoopNestBuilder(indexVarPtrs, indexVarLbs, indexVarUbs, indexVarSteps)([&] {
IndexedValue ivRes(result);
IndexedValue ivOperand(operand);
AffineIndexedValue ivRes(result);
AffineIndexedValue ivOperand(operand);
// On the LHS of the assignment, adjust the index for the concatenation axis.
llvm::SmallVector<ValueHandle, 5> resIndexHandles;
for (int i = 0; i < rank; i++)
{
resIndexHandles.push_back(i == concatenationAxis
? indexVars[i] + concatenationAxisPos
? indexVars[i] + ValueHandle(concatenationAxisPos)
: indexVars[i]);
}
......@@ -851,11 +847,11 @@ namespace
});
// Move up concatenation_axis_pos for the next operand.
concatenationAxisPos = concatenationAxisPos + vOperand.ub(concatenationAxis);
concatenationAxisPos =
ValueHandle(concatenationAxisPos) + vOperand.ub(concatenationAxis);
}
rewriter.replaceOp(op, {result});
return matchSuccess();
}
......@@ -874,14 +870,13 @@ namespace
auto axis = gatherOp.axis().getSExtValue();
// Create view to write into result.
MemRefView vRes(result), vParams(params), vIndices(indices);
MemRefBoundsCapture vRes(result), vParams(params), vIndices(indices);
// Indexed Values
IndexedValue iRes(result), iIndices(indices);
AffineIndexedValue iRes(result), iIndices(indices);
StdIndexedValue iParams(params);
// Construct outer loop for params dims. Exclude the axis dim.
SmallVector<ValueHandle, 4> paramsLbs, paramsUbs;
SmallVector<IndexHandle, 4> paramsIVs;
SmallVector<ValueHandle, 4> paramsLbs, paramsUbs, paramsIVs;
SmallVector<int64_t, 4> paramsSteps;
SmallVector<ValueHandle*, 4> paramsIVPtrs;
for (auto i = 0; i < vParams.rank(); i++)
......@@ -889,8 +884,8 @@ namespace
// skip gather axis
if (i == axis)
continue;
paramsLbs.push_back(IndexHandle(vParams.lb(i)));
paramsUbs.push_back(IndexHandle(vParams.ub(i)));
paramsLbs.push_back(vParams.lb(i));
paramsUbs.push_back(vParams.ub(i));
paramsSteps.push_back(vParams.step(i));
}
NGRAPH_CHECK(paramsLbs.size() == vParams.rank() - 1 &&
......@@ -898,17 +893,17 @@ namespace
paramsSteps.size() == paramsLbs.size(),
"Incorrect loop nest bounds size for gather params");
paramsIVs = makeIndexHandles(vParams.rank() - 1);
paramsIVPtrs = makeHandlePointers(MutableArrayRef<IndexHandle>(paramsIVs));
paramsIVs = ValueHandle::makeIndexHandles(vParams.rank() - 1);
paramsIVPtrs = makeHandlePointers(paramsIVs);
auto indicesLbs = vIndices.getLbs();
auto indicesUbs = vIndices.getUbs();
auto indicesSteps = vIndices.getSteps();
auto indicesIVs = makeIndexHandles(vIndices.rank());
auto indicesIVPtrs = makeHandlePointers(MutableArrayRef<IndexHandle>(indicesIVs));
auto indicesIVs = ValueHandle::makeIndexHandles(vIndices.rank());
auto indicesIVPtrs = makeHandlePointers(indicesIVs);
SmallVector<IndexHandle, 8> paramsIndices, resIndices;
SmallVector<ValueHandle, 8> paramsIndices, resIndices;
// Make sure we are going to create loops
NGRAPH_CHECK(vParams.rank() > 0, "Invalid size for indices steps");
......@@ -946,7 +941,7 @@ namespace
{
if (i == axis)
{
paramsIndices.push_back(IndexHandle(axisIdx));
paramsIndices.push_back(axisIdx);
}
else
{
......@@ -1022,10 +1017,10 @@ namespace
NGRAPH_CHECK(groups > 0, "Invalid number of groups");
// create outer group convolution loop
// for group = 0 to groups
IndexHandle iv;
ValueHandle lb = intrinsics::constant_index(0);
ValueHandle ub = intrinsics::constant_index(groups);
auto indexType = IndexType::get(rewriter.getContext());
ValueHandle iv(indexType);
ValueHandle lb = std_constant_index(0);
ValueHandle ub = std_constant_index(groups);
auto imagesType = images.getType().cast<MemRefType>();
auto filtersType = filters.getType().cast<MemRefType>();
......@@ -1043,13 +1038,13 @@ namespace
NGRAPH_CHECK(groupsInFilters || filtersShape[0] % groups == 0,
"Filters dim is not divisible by number of groups");
auto channelGroupSize = intrinsics::constant_index(imagesShape[1] / groups);
auto filtersGroupSize = intrinsics::constant_index(
groupsInFilters ? filtersShape[1] : filtersShape[0] / groups);
auto channelGroupSize = std_constant_index(imagesShape[1] / groups);
auto filtersGroupSize =
std_constant_index(groupsInFilters ? filtersShape[1] : filtersShape[0] / groups);
NGRAPH_CHECK(!groupsInFilters || groups == filtersShape[0]);
LoopBuilder::makeAffine(&iv, lb, ub, 1)([&] {
makeAffineLoopBuilder(&iv, lb, ub, 1)([&] {
// lower/upper bounds on image channel dim and kernels dim
auto cLb = iv * channelGroupSize;
auto cUb = cLb + channelGroupSize;
......@@ -1152,7 +1147,7 @@ namespace
castMemRef(inputs, outputs, rewriter, unrankedMemrefTy);
FuncOp callBackFunc = pass.getCallDecl(
"__mlir_callback_2_inputs",
"callback_2_inputs",
{unrankedMemrefTy, unrankedMemrefTy, unrankedMemrefTy, int64Ty, int64Ty},
{},
rewriter);
......@@ -1245,7 +1240,7 @@ namespace
auto int64Ty = rewriter.getIntegerType(64);
auto unrankedMemrefTy = UnrankedMemRefType::get(elemTy, 0);
auto callBackFunc = pass.getCallDecl(
"__mlir_callback_2_inputs",
"callback_2_inputs",
{unrankedMemrefTy, unrankedMemrefTy, unrankedMemrefTy, int64Ty, int64Ty},
{},
rewriter);
......@@ -1297,7 +1292,7 @@ namespace
elemTy == biasTy.getElementType(),
"Types mismatch in GemmOp");
MemRefView vRes(result), vLhs(lhs), vRhs(rhs), vBias(bias);
MemRefBoundsCapture vRes(result), vLhs(lhs), vRhs(rhs), vBias(bias);
NGRAPH_CHECK(vLhs.rank() == 2 && vRhs.rank() == 2 && vRes.rank() == 2 && vBias.rank() <= 2,
"Gemm operation is only supported for 2D tensors");
......@@ -1361,7 +1356,7 @@ namespace
auto int64Ty = rewriter.getIntegerType(64);
auto unrankedMemrefTy = UnrankedMemRefType::get(elemTy, 0);
auto callBackFunc = pass.getCallDecl("__mlir_callback_3_inputs",
auto callBackFunc = pass.getCallDecl("callback_3_inputs",
{unrankedMemrefTy,
unrankedMemrefTy,
unrankedMemrefTy,
......@@ -1425,7 +1420,7 @@ namespace
rewriter.getUnknownLoc(), static_cast<int64_t>(OpType::SOFTMAX), 64);
FuncOp callBackFunc =
pass.getCallDecl("__mlir_callback_1_input",
pass.getCallDecl("callback_1_input",
{unrankedMemrefTy, unrankedMemrefTy, int64Ty, int64Ty},
{},
rewriter);
......@@ -1511,11 +1506,12 @@ namespace
auto padBelow = padBelowAttr.getValue();
auto padAbove = padBelowAttr.getValue();
Type elemTy = images.getType().cast<MemRefType>().getElementType();
auto indexType = IndexType::get(rewriter.getContext());
// Create views
MemRefView vRes(result), vImages(images), vFilters(filters);
MemRefBoundsCapture vRes(result), vImages(images), vFilters(filters);
// Create indexed Values
IndexedValue iRes(result), iImages(images), iFilters(filters);
AffineIndexedValue iRes(result), iImages(images), iFilters(filters);
// Bounds on batch size N
ValueHandle batchLb = vImages.lb(0), batchUb = vImages.ub(0);
// Bounds on spatial dimensions
......@@ -1526,9 +1522,8 @@ namespace
unsigned spatialRank = vImages.rank() - 2;
// Result spatial indices and bounds
auto resSpatialIndices = makeIndexHandles(spatialRank);
auto resSpatialIndicesPtrs =
makeHandlePointers(MutableArrayRef<IndexHandle>(resSpatialIndices));
auto resSpatialIndices = ValueHandle::makeIndexHandles(spatialRank);
auto resSpatialIndicesPtrs = makeHandlePointers(resSpatialIndices);
SmallVector<int64_t, 4> resSteps, filtersSteps;
SmallVector<int, 4> padBelowIntValues;
bool withPadding = false;
......@@ -1610,9 +1605,8 @@ namespace
"Results spatial dims mismatches input");
// Filters spatial indices and bounds
auto filtersSpatialIndices = makeIndexHandles(spatialRank);
auto filtersSpatialIndicesPtrs =
makeHandlePointers(MutableArrayRef<IndexHandle>(filtersSpatialIndices));
auto filtersSpatialIndices = ValueHandle::makeIndexHandles(spatialRank);
auto filtersSpatialIndicesPtrs = makeHandlePointers(filtersSpatialIndices);
for (auto i = 0; i < spatialRank; i++)
{
......@@ -1658,23 +1652,22 @@ namespace
// Initialize output to zero
{
IndexHandle n, k, c;
auto resSpatialIndices = makeIndexHandles(spatialRank);
auto resSpatialIndicesPtrs =
makeHandlePointers(MutableArrayRef<IndexHandle>(resSpatialIndices));
ValueHandle n(indexType), k(indexType), c(indexType);
auto resSpatialIndices = ValueHandle::makeIndexHandles(spatialRank);
auto resSpatialIndicesPtrs = makeHandlePointers(resSpatialIndices);
LoopBuilder::makeAffine(&n, batchLb, batchUb, 1)([&] {
LoopBuilder::makeAffine(&k, numFiltersLb, numFiltersUb, 1)([&] {
makeAffineLoopBuilder(&n, batchLb, batchUb, 1)([&] {
makeAffineLoopBuilder(&k, numFiltersLb, numFiltersUb, 1)([&] {
AffineLoopNestBuilder(
resSpatialIndicesPtrs, resSpatialLbs, resSpatialUbs, resSteps)([&] {
SmallVector<IndexHandle, 4> resIndices;
SmallVector<ValueHandle, 4> resIndices;
// Result indices
resIndices.push_back(n);
if (groupConvolution && groupsInFilters)
{
// compute global C_OUT from gID and k
// gId * C_OUT (num of filters) + k
resIndices.push_back(IndexHandle(ValueHandle(gId) * numFiltersUb + k));
resIndices.push_back(ValueHandle(gId) * numFiltersUb + k);
}
else
{
......@@ -1689,31 +1682,31 @@ namespace
});
}
IndexHandle n, k, c;
ValueHandle n(indexType), k(indexType), c(indexType);
// Convolution loop
LoopBuilder::makeAffine(&n, batchLb, batchUb, 1)([&] {
makeAffineLoopBuilder(&n, batchLb, batchUb, 1)([&] {
// Number of filters loop
LoopBuilder::makeAffine(&k, numFiltersLb, numFiltersUb, 1)([&] {
makeAffineLoopBuilder(&k, numFiltersLb, numFiltersUb, 1)([&] {
// Channels loop
LoopBuilder::makeAffine(&c, numChannelsLb, numChannelsUb, 1)([&] {
makeAffineLoopBuilder(&c, numChannelsLb, numChannelsUb, 1)([&] {
// Results loop
AffineLoopNestBuilder(
resSpatialIndicesPtrs, resSpatialLbs, resSpatialUbs, resSteps)([&] {
// Compute image start indices
SmallVector<IndexHandle, 4> imgStartIndices;
SmallVector<ValueHandle, 4> imgStartIndices;
for (auto i = 0; i < spatialRank; i++)
{
IntegerAttr iAttr = strides[i].cast<IntegerAttr>();
auto stride = intrinsics::constant_index(iAttr.getInt());
imgStartIndices.push_back(IndexHandle(resSpatialIndices[i] * stride));
auto stride = std_constant_index(iAttr.getInt());
imgStartIndices.push_back(resSpatialIndices[i] * stride);
}
SmallVector<IndexHandle, 4> resIndices;
SmallVector<ValueHandle, 4> resIndices;
// Result indices
resIndices.push_back(n);
if (groupConvolution && groupsInFilters)
{
// gId * C_OUT (num of filters) + k
resIndices.push_back(IndexHandle(ValueHandle(gId) * numFiltersUb + k));
resIndices.push_back(ValueHandle(gId) * numFiltersUb + k);
}
else
{
......@@ -1727,15 +1720,14 @@ namespace
filtersSpatialLbs,
filtersSpatialUbs,
filtersSteps)([&] {
SmallVector<IndexHandle, 4> imgIndices, filtersIndices;
SmallVector<ValueHandle, 4> imgIndices, filtersIndices;
// Image indices
// Here we compute the virtual start index into the padded image.
imgIndices.push_back(n);
imgIndices.push_back(c);
for (auto i = 0; i < spatialRank; i++)
{
imgIndices.push_back(
IndexHandle(imgStartIndices[i] + filtersSpatialIndices[i]));
imgIndices.push_back(imgStartIndices[i] + filtersSpatialIndices[i]);
}
// Filter indices
......@@ -1744,14 +1736,14 @@ namespace
// index
if (groupConvolution && groupsInFilters)
{
filtersIndices.push_back(IndexHandle(gId));
filtersIndices.push_back(ValueHandle(gId));
}
filtersIndices.push_back(k);
// subtract lower bound of channel
// if we are doing group convolution this bound will advance based
// on the group id. For the filters, it should always start from 0
filtersIndices.push_back(IndexHandle(c - numChannelsLb));
filtersIndices.push_back(c - numChannelsLb);
filtersIndices.insert(filtersIndices.end(),
filtersSpatialIndices.begin(),
filtersSpatialIndices.end());
......@@ -1759,7 +1751,7 @@ namespace
if (withPadding)
{
// if args : img dims, img lbs, img ubs
SmallVector<IndexHandle, 4>::iterator it = imgIndices.begin();
SmallVector<ValueHandle, 4>::iterator it = imgIndices.begin();
std::advance(it, 2);
SmallVector<Value, 4> affineIfArgs(it, imgIndices.end());
affineIfArgs.insert(
......@@ -1777,14 +1769,14 @@ namespace
ScopedContext scope(rewriter, loc);
// We must subtract pad below before img load, since the
// physical image is not padded
SmallVector<IndexHandle, 4> adjustedImgIndices;
SmallVector<ValueHandle, 4> adjustedImgIndices;
adjustedImgIndices.push_back(n);
adjustedImgIndices.push_back(c);
for (auto i = 0; i < spatialRank; i++)
{
adjustedImgIndices.push_back(IndexHandle(
adjustedImgIndices.push_back(
imgIndices[2 + i] -
intrinsics::constant_index(padBelowIntValues[i])));
std_constant_index(padBelowIntValues[i]));
}
iRes(resIndices) =
iRes(resIndices) +
......@@ -1821,15 +1813,15 @@ namespace
ScopedContext scope(rewriter, loc);
// Views
MemRefView vRes(result), vLHS(lhs);
MemRefBoundsCapture vRes(result), vLHS(lhs);
// Index Values
IndexedValue iRes(result), iLHS(lhs);
AffineIndexedValue iRes(result), iLHS(lhs);
// Bounds Index Handles
auto lbs = vLHS.getLbs();
auto ubs = vLHS.getUbs();
// Loop induction vars
auto ivs = makeIndexHandles(vLHS.rank());
auto pivs = makeHandlePointers(MutableArrayRef<IndexHandle>(ivs));
auto ivs = ValueHandle::makeIndexHandles(vLHS.rank());
auto pivs = makeHandlePointers(ivs);
// Steps
auto steps = vLHS.getSteps();
......@@ -1867,15 +1859,15 @@ namespace
ScopedContext scope(rewriter, loc);
// Views
MemRefView vRes(result), vLHS(lhs), vRHS(rhs);
MemRefBoundsCapture vRes(result), vLHS(lhs), vRHS(rhs);
// Index Values
IndexedValue iRes(result), iLHS(lhs), iRHS(rhs);
AffineIndexedValue iRes(result), iLHS(lhs), iRHS(rhs);
// Bounds Index Handles
auto lbs = vLHS.getLbs();
auto ubs = vLHS.getUbs();
// Loop induction vars
auto ivs = makeIndexHandles(vLHS.rank());
auto pivs = makeHandlePointers(MutableArrayRef<IndexHandle>(ivs));
auto ivs = ValueHandle::makeIndexHandles(vLHS.rank());
auto pivs = makeHandlePointers(ivs);
// Steps
auto steps = vLHS.getSteps();
// element type of the operand
......@@ -1900,65 +1892,57 @@ namespace
iRes(ivs) = iLHS(ivs) / iRHS(ivs);
}
// TODO(pthoreho) For all comparision operators, use
// edsc::intrinsics::zero_extendi(ValueHandle(iLHS(ivs)) !=
// zero_extendi(ValueHandle(iLHS(ivs)) !=
// ValueHandle(iRHS(ivs)), IntegerType::get(8, op->getContext()));
// instead of edsc::intrinsics::select once `zero_extendi` is
// instead of std_select once `zero_extendi` is
// made available in the edsc::intrinsics namescope in MLIR repo.
else if (isa<NGGreaterOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) > ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) > ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
}
else if (isa<NGLessOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) < ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) < ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
}
else if (isa<NGGreaterEqOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) >= ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) >= ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
}
else if (isa<NGLessEqOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) <= ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) <= ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
}
else if (isa<NGEqOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) == ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) == ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
}
else if (isa<NGNotEqOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) != ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) != ValueHandle(iRHS(ivs)),
createOneConstant(elemTy),
createZeroConstant(elemTy));
}
else if (isa<NGMaxOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) > ValueHandle(iRHS(ivs)),
ValueHandle(iLHS(ivs)),
ValueHandle(iRHS(ivs)));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) > ValueHandle(iRHS(ivs)),
ValueHandle(iLHS(ivs)),
ValueHandle(iRHS(ivs)));
}
else if (isa<NGMinOp>(op))
{
iRes(ivs) =
edsc::intrinsics::select(ValueHandle(iLHS(ivs)) < ValueHandle(iRHS(ivs)),
ValueHandle(iLHS(ivs)),
ValueHandle(iRHS(ivs)));
iRes(ivs) = std_select(ValueHandle(iLHS(ivs)) < ValueHandle(iRHS(ivs)),
ValueHandle(iLHS(ivs)),
ValueHandle(iRHS(ivs)));
}
else
{
......@@ -1995,10 +1979,10 @@ namespace
Value result = pass.buildOutputDefs(op, rewriter)[0];
// Views
MemRefView vRes(result), vArg(arg);
MemRefBoundsCapture vRes(result), vArg(arg);
// Index Values
StdIndexedValue iRes(result), stdArg(arg);
IndexedValue affineArg(arg);
AffineIndexedValue affineArg(arg);
// Bounds Index Handles
auto resLbs = vRes.getLbs();
auto resUbs = vRes.getUbs();
......@@ -2008,8 +1992,8 @@ namespace
Type resTy = result.getType().cast<MemRefType>().getElementType();
// Generate loop nest that initializes result to lower bound of the axis to be reduced.
{
auto ivs = makeIndexHandles(vRes.rank());
auto pivs = makeHandlePointers(MutableArrayRef<IndexHandle>(ivs));
auto ivs = ValueHandle::makeIndexHandles(vRes.rank());
auto pivs = makeHandlePointers(ivs);
auto steps = vRes.getSteps();
auto initVal = vArg.lb(axis);
AffineLoopNestBuilder(pivs, resLbs, resUbs, steps)(
......@@ -2018,10 +2002,10 @@ namespace
// Generate loop nest that computes the actual index reduction.
{
auto allIVs = makeIndexHandles(vArg.rank());
auto pAllIVs = makeHandlePointers(MutableArrayRef<IndexHandle>(allIVs));
auto allIVs = ValueHandle::makeIndexHandles(vArg.rank());
auto pAllIVs = makeHandlePointers(allIVs);
auto steps = vArg.getSteps();
SmallVector<IndexHandle, 8> nonRedIVs;
SmallVector<ValueHandle, 8> nonRedIVs;
Type resTy = result.getType().cast<MemRefType>().getElementType();
NGRAPH_CHECK(resTy.isa<IntegerType>(),
......@@ -2049,10 +2033,8 @@ namespace
// Select the min/max value and cast it back to integer type before storing it.
ValueHandle newRedIdx =
std::is_same<RedOp, NGArgMinRedOp>()
? edsc::intrinsics::select(
affineArg(allIVs) < stdArg(tempIVs), allIVs[axis], currRedIdx)
: edsc::intrinsics::select(
stdArg(tempIVs) < affineArg(allIVs), allIVs[axis], currRedIdx);
? std_select(affineArg(allIVs) < stdArg(tempIVs), allIVs[axis], currRedIdx)
: std_select(stdArg(tempIVs) < affineArg(allIVs), allIVs[axis], currRedIdx);
iRes(nonRedIVs) = ValueHandle::create<IndexCastOp>(newRedIdx, resTy);
});
......@@ -2123,7 +2105,7 @@ namespace
castMemRef(inputs, outputs, rewriter, unrankedMemrefTy);
FuncOp callBackFunc =
pass.getCallDecl("__mlir_callback_1_input",
pass.getCallDecl("callback_1_input",
{unrankedMemrefTy, unrankedMemrefTy, int64Ty, int64Ty},
{},
rewriter);
......@@ -2168,11 +2150,11 @@ namespace
{
if (floatTy.isF32())
{
return intrinsics::constant_float(llvm::APFloat(0.0f), floatTy);
return std_constant_float(llvm::APFloat(0.0f), floatTy);
}
else if (floatTy.isF64())
{
return intrinsics::constant_float(llvm::APFloat(0.0), floatTy);
return std_constant_float(llvm::APFloat(0.0), floatTy);
}
else
{
......@@ -2181,7 +2163,7 @@ namespace
}
else if (auto intTy = type.dyn_cast<IntegerType>())
{
return intrinsics::constant_int(0, intTy.getWidth());
return std_constant_int(0, intTy.getWidth());
}
NGRAPH_UNREACHABLE("Unsupported type");
}
......@@ -2192,11 +2174,11 @@ namespace
{
if (floatTy.isF32())
{
return intrinsics::constant_float(llvm::APFloat(1.0f), floatTy);
return std_constant_float(llvm::APFloat(1.0f), floatTy);
}
else if (floatTy.isF64())
{
return intrinsics::constant_float(llvm::APFloat(1.0f), floatTy);
return std_constant_float(llvm::APFloat(1.0f), floatTy);
}
else
{
......@@ -2205,7 +2187,7 @@ namespace
}
else if (auto intTy = type.dyn_cast<IntegerType>())
{
return intrinsics::constant_int(1, intTy.getWidth());
return std_constant_int(1, intTy.getWidth());
}
NGRAPH_UNREACHABLE("Unsupported type");
}
......
src/contrib/mlir/core/pass/ng_dialect_fused_ops.cpp
View file @ 310fcf07
......@@ -23,9 +23,7 @@
#include "contrib/mlir/core/ngraph_dialect/type.hpp"
#include <llvm/IR/Module.h>
#include <mlir/EDSC/Builders.h>
#include <mlir/EDSC/Helpers.h>
#include <mlir/EDSC/Intrinsics.h>
#include <mlir/Dialect/AffineOps/EDSC/Builders.h>
#include <mlir/IR/IntegerSet.h>
#include <mlir/IR/MLIRContext.h>
#include <mlir/IR/StandardTypes.h>
......
src/contrib/mlir/runtime/cpu/cpu_callbacks.cpp
View file @ 310fcf07
......@@ -719,7 +719,7 @@ static void __mlir_cblas_sgemm_with_bias(StaticMemRef* memRefmatA,
}
}
extern "C" void __mlir_callback_1_input(void* input, void* output, size_t index, OpType type)
extern "C" void _mlir_ciface_callback_1_input(void* input, void* output, size_t index, OpType type)
{
auto unrankedMemRefInput = reinterpret_cast<UnrankedMemRef*>(input);
auto unrankedMemRefOutput = reinterpret_cast<UnrankedMemRef*>(output);
......@@ -752,8 +752,8 @@ extern "C" void __mlir_callback_1_input(void* input, void* output, size_t index,
}
}
extern "C" void
__mlir_callback_2_inputs(void* input0, void* input1, void* output, size_t index, OpType type)
extern "C" void _mlir_ciface_callback_2_inputs(
void* input0, void* input1, void* output, size_t index, OpType type)
{
auto unrankedMemRefInput0 = reinterpret_cast<UnrankedMemRef*>(input0);
auto unrankedMemRefInput1 = reinterpret_cast<UnrankedMemRef*>(input1);
......@@ -780,7 +780,7 @@ extern "C" void
}
}
extern "C" void __mlir_callback_3_inputs(
extern "C" void _mlir_ciface_callback_3_inputs(
void* input0, void* input1, void* input2, void* output, size_t index, OpType type)
{
auto unrankedMemRefInput0 = reinterpret_cast<UnrankedMemRef*>(input0);
......
src/contrib/mlir/runtime/cpu/cpu_runtime.cpp
View file @ 310fcf07
......@@ -83,7 +83,7 @@ void MLIRCPURuntime::bindArguments(const std::vector<MemRefArg>& args)
{
NGRAPH_CHECK(m_module, "MLIR module is not ready.");
auto func = m_module->lookupSymbol<mlir::LLVM::LLVMFuncOp>("main");
auto func = m_module->lookupSymbol<mlir::LLVM::LLVMFuncOp>("_mlir_ciface_main");
NGRAPH_CHECK(func && !func.getBlocks().empty(), "Function not found");
// Set external arguments
......@@ -127,14 +127,15 @@ void MLIRCPURuntime::execute()
// uniformity reasons, it takes a list of type-erased pointers to arguments.
// Please, note that 'invoke' method is overloaded with a parameter pack version.
// Make sure the MutableArrayRef version is invoked.
auto invocationResult = m_engine->invoke("main", llvm::MutableArrayRef<void*>(m_invokeArgs));
auto invocationResult =
m_engine->invoke("_mlir_ciface_main", llvm::MutableArrayRef<void*>(m_invokeArgs));
if (clDumpObjectFile)
{
m_engine->dumpToObjectFile(clObjectFilename.empty() ? "jitted_mlir.o"
: clObjectFilename.getValue());
}
NGRAPH_CHECK(!invocationResult, "JIT invocation of 'main' failed\n");
NGRAPH_CHECK(!invocationResult, "JIT invocation of '_mlir_ciface_main' failed\n");
}
void MLIRCPURuntime::cleanup()
......
src/contrib/mlir/tools/ngraph-opt/CMakeLists.txt
View file @ 310fcf07
......@@ -16,7 +16,7 @@
set(LIBS
mlir_backend
MLIROptMain
MLIROptLib
MLIRPass
MLIRParser
LLVMSupport
......
src/contrib/mlir/utils.cpp
View file @ 310fcf07
......@@ -21,10 +21,21 @@
#include "contrib/mlir/core/ngraph_dialect/dialect.hpp"
#include <llvm/Support/CommandLine.h>
#include <llvm/Support/Debug.h>
#include <mlir/Dialect/AffineOps/AffineOps.h>
#include <mlir/Dialect/LLVMIR/LLVMDialect.h>
#include <mlir/Dialect/LoopOps/LoopOps.h>
#include <mlir/Dialect/StandardOps/Ops.h>
#include <mlir/Dialect/VectorOps/VectorOps.h>
#include <mlir/IR/Dialect.h>
#include <mlir/IR/MLIRContext.h>
#include <mlir/Pass/Pass.h>
#include <mlir/Transforms/LocationSnapshot.h>
#include <mlir/Transforms/Passes.h>
#include <llvm/Support/CommandLine.h>
#include <llvm/Support/Debug.h>
using namespace mlir;
static llvm::cl::opt<bool> clPrintIRAfterAll(
"ngraph-print-ir-after-all",
......@@ -35,15 +46,47 @@ static llvm::cl::opt<bool> clPrintIRAfterAll(
void ngraph::runtime::ngmlir::initializeNGraphMLIR()
{
// Initialize a dialect only once.
// We currently have no way to query if a dialect is previously
// registered. So using a global flag instead.
static bool init = false;
if (!init)
{
mlir::registerDialect<mlir::NGraphOpsDialect>();
init = true;
}
// Initialize MLIR dialects and passes only once.
static bool init_once = []() {
// In-tree Dialects.
registerDialect<AffineOpsDialect>();
registerDialect<LLVM::LLVMDialect>();
registerDialect<loop::LoopOpsDialect>();
registerDialect<StandardOpsDialect>();
registerDialect<vector::VectorOpsDialect>();
// nGraph dialects.
registerDialect<mlir::NGraphOpsDialect>();
// In-tree passes.
// No-op to avoid DCE on the following pass initializations.
if (std::getenv("bar") != (char*)-1)
return false;
createCanonicalizerPass();
createCSEPass();
createVectorizePass({});
createLoopUnrollPass();
createLoopUnrollAndJamPass();
createSimplifyAffineStructuresPass();
createLoopFusionPass();
createLoopInvariantCodeMotionPass();
createAffineLoopInvariantCodeMotionPass();
createPipelineDataTransferPass();
createLowerAffinePass();
createLoopTilingPass(0);
createLoopCoalescingPass();
createAffineDataCopyGenerationPass(0, 0);
createMemRefDataFlowOptPass();
createStripDebugInfoPass();
createPrintOpStatsPass();
createInlinerPass();
createSymbolDCEPass();
createLocationSnapshotPass({});
return true;
}();
(void)init_once;
}
void ngraph::runtime::ngmlir::dumpMlirModule(const std::string msg, mlir::ModuleOp module)
......
src/ngraph/frontend/onnx_import/CMakeLists.txt
View file @ 310fcf07
......@@ -171,6 +171,8 @@ add_library(onnx_import STATIC
op/reshape.hpp
op/reverse_sequence.cpp
op/reverse_sequence.hpp
op/round.cpp
op/round.hpp
op/scatter_nd.cpp
op/scatter_nd.hpp
op/selu.cpp
......
src/ngraph/frontend/onnx_import/op/average_pool.cpp
View file @ 310fcf07
......@@ -16,7 +16,6 @@
#include "average_pool.hpp"
#include "ngraph/node.hpp"
#include "ngraph/op/avg_pool.hpp"
#include "utils/pooling_factory.hpp"
namespace ngraph
......@@ -29,7 +28,7 @@ namespace ngraph
{
NodeVector average_pool(const Node& node)
{
return pooling::PoolingFactory(node).make_avg_pool();
return pooling::LocalPoolingFactory(node).make_avg_pool();
}
} // namespace set_1
......
src/ngraph/frontend/onnx_import/op/max_pool.cpp
View file @ 310fcf07
......@@ -31,7 +31,7 @@ namespace ngraph
{
NodeVector max_pool(const Node& node)
{
auto max_pool = pooling::PoolingFactory(node).make_max_pool();
auto max_pool = pooling::LocalPoolingFactory(node).make_max_pool();
max_pool.emplace_back(std::make_shared<NullNode>()); // Indices (optional)
return max_pool;
}
......
src/ngraph/frontend/onnx_import/op/onehot.cpp
View file @ 310fcf07
......@@ -42,9 +42,9 @@ namespace ngraph
auto off_on_values =
std::make_shared<default_opset::Split>(values, split_axis, 2);
auto off_value =
reshape::interpret_as_scalar(get_output_element(off_on_values, 0ul));
reshape::interpret_as_scalar(get_output_element(off_on_values, size_t{0}));
auto on_value =
reshape::interpret_as_scalar(get_output_element(off_on_values, 1ul));
reshape::interpret_as_scalar(get_output_element(off_on_values, size_t{1}));
auto axis = node.get_attribute_value<std::int64_t>("axis", -1);
......
src/ngraph/frontend/onnx_import/op/pad.cpp
View file @ 310fcf07
......@@ -65,14 +65,19 @@ namespace ngraph
NodeVector pad(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
const Shape& data_shape = data->get_shape();
const auto data_rank =
node.get_ng_inputs().at(0)->get_output_partial_shape(0).rank();
CHECK_VALID_NODE(
node, data_rank.is_static(), "Data rank must be static for pad op");
const auto data_rank_value = static_cast<size_t>(data_rank);
double value = node.get_attribute_value<double>("value", 0);
const std::string mode =
node.get_attribute_value<std::string>("mode", "constant");
ngraph::op::PadMode pad_mode = get_pad_mode(mode);
auto paddings = convpool::get_pads(node, data_shape);
const auto paddings = convpool::get_pads(node, data_rank_value);
ngraph::CoordinateDiff padding_below = paddings.first;
ngraph::CoordinateDiff padding_above = paddings.second;
......
src/ngraph/frontend/onnx_import/op/round.cpp 0 → 100644
View file @ 310fcf07
//*****************************************************************************
// Copyright 2017-2020 Intel Corporation
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//*****************************************************************************
#include <memory>
#include "ngraph/opsets/opset0.hpp"
#include "round.hpp"
namespace ngraph
{
namespace onnx_import
{
namespace op
{
namespace set_1
{
NodeVector round(const Node& node)
{
const std::shared_ptr<ngraph::Node> data{node.get_ng_inputs().at(0)};
return {std::make_shared<ngraph::opset0::Round>(data)};
}
} // namespace set_1
} // namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/round.hpp 0 → 100644
View file @ 310fcf07
//*****************************************************************************
// Copyright 2017-2020 Intel Corporation
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//*****************************************************************************
#pragma once
#include "core/node.hpp"
#include "ngraph/node.hpp"
namespace ngraph
{
namespace onnx_import
{
namespace op
{
namespace set_1
{
NodeVector round(const Node& node);
} // namespace set_1
} // namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/ops_bridge.cpp
View file @ 310fcf07
......@@ -101,6 +101,7 @@
#include "op/relu.hpp"
#include "op/reshape.hpp"
#include "op/reverse_sequence.hpp"
#include "op/round.hpp"
#include "op/scatter_nd.hpp"
#include "op/selu.hpp"
#include "op/shape.hpp"
......@@ -334,6 +335,7 @@ namespace ngraph
REGISTER_OPERATOR("Relu", 1, relu);
REGISTER_OPERATOR("Reshape", 1, reshape);
REGISTER_OPERATOR("ReverseSequence", 1, reverse_sequence);
REGISTER_OPERATOR("Round", 1, round);
REGISTER_OPERATOR("ScatterND", 1, scatter_nd);
REGISTER_OPERATOR("Selu", 1, selu);
REGISTER_OPERATOR("Shape", 1, shape);
......
src/ngraph/frontend/onnx_import/utils/convpool.cpp
View file @ 310fcf07
......@@ -38,28 +38,41 @@ namespace ngraph
namespace detail
{
Strides get_strides_helper(const Node& node,
const std::string& name,
const Shape& kernel_shape)
/// \brief Helper method used to read vector attribute
/// \note Default value is vector of size spatial dims filled with
/// ones
///
/// \param node Node from which attribute is read
/// \param attr_name Attribute name (such as `strides`, `dilations`)
///
/// \return Read vector attribute if available or default value
std::vector<std::size_t> get_attribute_value(const Node& node,
const std::string& attr_name)
{
return node.get_attribute_value<std::vector<std::size_t>>(
name, std::vector<std::size_t>(kernel_shape.size(), 1UL));
if (node.has_attribute(attr_name))
{
return node.get_attribute_value<std::vector<std::size_t>>(attr_name);
}
const auto data_rank =
node.get_ng_inputs().at(0)->get_output_partial_shape(0).rank();
CHECK_VALID_NODE(node,
data_rank.is_static(),
"If '",
attr_name,
"' is not provided data rank must be static");
const auto data_spatial_dims = static_cast<size_t>(data_rank) - 2;
return std::vector<std::size_t>(data_spatial_dims, 1UL);
}
} // namespace detail
Strides get_strides(const Node& node, const Shape& kernel_shape)
{
return detail::get_strides_helper(node, "strides", kernel_shape);
}
Strides get_strides(const Node& node)
{
return get_strides(node, get_kernel_shape(node));
return detail::get_attribute_value(node, "strides");
}
Strides get_dilations(const Node& node)
{
return detail::get_strides_helper(node, "dilations", get_kernel_shape(node));
return detail::get_attribute_value(node, "dilations");
}
ngraph::op::PadType get_auto_pad(const Node& node)
......@@ -90,16 +103,16 @@ namespace ngraph
}
std::pair<CoordinateDiff, CoordinateDiff> get_pads(const Node& node,
const Shape& kernel_shape)
const size_t kernel_rank)
{
CoordinateDiff pads(kernel_shape.size(), 0);
CoordinateDiff pads(kernel_rank, 0);
if (node.has_attribute("pads"))
{
auto pads_int64 = node.get_attribute_value<std::vector<int64_t>>("pads");
pads = CoordinateDiff{std::begin(pads_int64), std::end(pads_int64)};
}
if (pads.size() == kernel_shape.size() * 2)
if (pads.size() == kernel_rank * 2)
{
return {{std::begin(pads), std::begin(pads) + pads.size() / 2},
{std::begin(pads) + pads.size() / 2, std::end(pads)}};
......@@ -112,6 +125,18 @@ namespace ngraph
}
}
std::pair<CoordinateDiff, CoordinateDiff> get_pads(const Node& node)
{
const auto data_rank =
node.get_ng_inputs().at(0)->get_output_partial_shape(0).rank();
CHECK_VALID_NODE(node,
data_rank.is_static(),
"The rank of node must be static in order to calculate pads");
const auto data_spatial_dims = static_cast<size_t>(data_rank) - 2;
return get_pads(node, data_spatial_dims);
}
void calculate_auto_pads(const Shape& data_shape,
const Shape& filter_shape,
const Strides& strides,
......
src/ngraph/frontend/onnx_import/utils/convpool.hpp
View file @ 310fcf07
......@@ -33,13 +33,6 @@ namespace ngraph
/// \return The kernel Shape object representing its dimensions (height, width, depth).
Shape get_kernel_shape(const Node& node);
/// \brief Get number of pixels to stride operation by in each direction.
///
/// \param node The Node ptr representing Conv or Pool operation.
/// \param kernel_shape The shape of the kernel which we retrieve strides for.
/// \return The kernel Shape object representing its dimensions (height, width, depth).
Strides get_strides(const Node& node, const Shape& kernel_shape);
/// \brief Get number of pixels to stride operation by in each direction.
///
/// \param node The Node ptr representing Conv or Pool operation.
......@@ -59,12 +52,12 @@ namespace ngraph
/// `pads` value should follow [x1_begin, x2_begin..., x1_end, x2_end,...].
///
/// \param node The Node ptr representing ONNX operation.
/// \param kernel_shape The shape of the kernel which we retrieve pads for.
/// \param kernel_rank The rank of the kernel which we retrieve pads for.
///
/// \return A pair of (padding_above, padding_below), which elements contains number of
/// pixels to pad in respective dimensions (height, width, depth).
std::pair<CoordinateDiff, CoordinateDiff> get_pads(const Node& node,
const Shape& kernel_shape);
const size_t kernel_rank);
/// \brief Get padding values for the operation described by an ONNX node.
/// \details Values are taken from the `pads` attribute.
......@@ -75,11 +68,7 @@ namespace ngraph
///
/// \return A pair of (padding_above, padding_below), which elements contains number of
/// pixels to pad in respective dimensions (height, width, depth).
inline std::pair<CoordinateDiff, CoordinateDiff> get_pads(const Node& node)
{
return get_pads(node, get_kernel_shape(node));
}
std::pair<CoordinateDiff, CoordinateDiff> get_pads(const Node& node);
///
/// \brief Calculate paddings with respect to auto_pad value.
......
src/ngraph/frontend/onnx_import/utils/pooling_factory.cpp
View file @ 310fcf07
......@@ -17,6 +17,7 @@
#include <iterator>
#include "default_opset.hpp"
#include "exceptions.hpp"
#include "ngraph/coordinate_diff.hpp"
#include "utils/convpool.hpp"
#include "utils/pooling_factory.hpp"
......@@ -30,12 +31,11 @@ namespace ngraph
PoolingFactory::PoolingFactory(const Node& node)
: m_onnx_node{node}
, m_inputs{node.get_ng_inputs()}
, m_kernel_shape{convpool::get_kernel_shape(node)}
, m_strides{convpool::get_strides(node)}
, m_dilations{convpool::get_dilations(node)}
, m_auto_pad{convpool::get_auto_pad(node)}
{
auto paddings = convpool::get_pads(node);
const auto paddings = convpool::get_pads(node);
const CoordinateDiff& padding_above{paddings.second};
const CoordinateDiff& padding_below{paddings.first};
m_padding_below = Shape{std::begin(padding_below), std::end(padding_below)};
......@@ -44,7 +44,7 @@ namespace ngraph
NodeVector PoolingFactory::make_avg_pool() const
{
bool count_include_pad =
const bool count_include_pad =
m_onnx_node.get_attribute_value<std::int64_t>("count_include_pad", 0);
return {std::make_shared<default_opset::AvgPool>(m_inputs.at(0),
m_strides,
......@@ -67,13 +67,31 @@ namespace ngraph
m_auto_pad)};
}
LocalPoolingFactory::LocalPoolingFactory(const Node& node)
: PoolingFactory(node)
{
// Kernel shape is required
m_kernel_shape = node.get_attribute_value<std::vector<std::size_t>>("kernel_shape");
}
GlobalPoolingFactory::GlobalPoolingFactory(const Node& node)
: PoolingFactory(node)
{
// Correct the kernel shape.
const Shape& data_shape{m_inputs.at(0)->get_shape()};
const auto data_shape = node.get_ng_inputs().at(0)->get_output_partial_shape(0);
const auto data_rank = data_shape.rank();
CHECK_VALID_NODE(
node, data_rank.is_static(), "Data rank must be static for global pooling ops");
Shape kernel_shape;
for (auto i = 2; i < static_cast<size_t>(data_rank); ++i)
{
CHECK_VALID_NODE(node,
data_shape[i].is_static(),
"All spatial dimensions must be known for global pooling ops");
kernel_shape.emplace_back(static_cast<size_t>(data_shape[i]));
}
// Set shape to all but {N,C} axes.
m_kernel_shape = Shape{std::next(std::begin(data_shape), 2), std::end(data_shape)};
m_kernel_shape = kernel_shape;
}
} // namespace pooling
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/utils/pooling_factory.hpp
View file @ 310fcf07
......@@ -48,7 +48,6 @@ namespace ngraph
class PoolingFactory
{
public:
explicit PoolingFactory(const Node& node);
virtual ~PoolingFactory() = default;
///
......@@ -64,6 +63,8 @@ namespace ngraph
NodeVector make_max_pool() const;
protected:
explicit PoolingFactory(const Node& node);
Node m_onnx_node;
const NodeVector m_inputs;
Shape m_kernel_shape;
......@@ -75,9 +76,20 @@ namespace ngraph
};
///
/// \brief Factory class which generates sub-graphs for ONNX 'global' pooling
/// \brief Factory class which generates sub-graphs for ONNX 'local' pooling
/// operators.
/// \note Kernel shape attribute is required
class LocalPoolingFactory : public PoolingFactory
{
public:
explicit LocalPoolingFactory(const Node& node);
virtual ~LocalPoolingFactory() = default;
};
///
/// \brief Factory class which generates sub-graphs for ONNX 'global' pooling
/// operators.
/// \note Kernel shape is calculated based on spatial dims
class GlobalPoolingFactory : public PoolingFactory
{
public:
......
src/ngraph/op/gather.cpp
View file @ 310fcf07
......@@ -130,7 +130,7 @@ void op::v1::Gather::validate_and_infer_types()
").");
}
auto axis = get_axis();
int64_t axis = get_axis();
if (input_rank.is_static() && axis != AXIS_NOT_SET_VALUE)
{
NODE_VALIDATION_CHECK(this,
......
src/ngraph/pattern/matcher.cpp
View file @ 310fcf07
......@@ -40,10 +40,18 @@ namespace ngraph
{
if (m_restore)
{
m_matcher->m_matched_list.erase(m_matcher->m_matched_list.begin() + m_watermark,
m_matcher->m_matched_list.end());
m_matcher->m_pattern_value_maps.erase(m_pattern_value_maps.begin() + m_capture_size,
m_pattern_value_maps.end());
if (!m_matcher->m_matched_list.empty())
{
m_matcher->m_matched_list.erase(m_matcher->m_matched_list.begin() + m_watermark,
m_matcher->m_matched_list.end());
}
if (!m_pattern_value_maps.empty())
{
m_matcher->m_pattern_value_maps.erase(
m_pattern_value_maps.begin() + m_capture_size, m_pattern_value_maps.end());
}
m_matcher->m_pattern_map = m_pattern_value_map;
}
}
......
src/ngraph/runtime/cpu/cpu_runtime_context.hpp
View file @ 310fcf07
......@@ -36,7 +36,7 @@
namespace mkldnn
{
class primitive;
struct primitive;
}
namespace ngraph
......
src/ngraph/runtime/cpu/pass/cpu_mkldnn_primitive_build.hpp
View file @ 310fcf07
......@@ -35,7 +35,7 @@
namespace mkldnn
{
class primitive;
struct primitive;
}
namespace ngraph
......
src/ngraph/runtime/gpu/unit_test.manifest
View file @ 310fcf07
......@@ -453,6 +453,7 @@ model_gatherND_int32
model_gatherND_float
model_pad_constant
model_reciprocal
model_round
tile_3d_small_data_rank
tile_3d_few_repeats
select_v1
......
src/ngraph/runtime/plaidml/unit_test.manifest
View file @ 310fcf07
......@@ -282,6 +282,7 @@ model_argmax_int32
model_argmin_int32
model_lp_norm_default
model_instance_normalization
model_round
# passing locally, fails closeness checks in CI which may be too strict
elu
......
test/mlir/affine_conversion/callback_ops.mlir
View file @ 310fcf07
......@@ -10,7 +10,7 @@
// CHECK: %[[C2:.*]] = constant {{[0-9]+}} : i64
// CHECK: %0 = memref_cast %arg0 : memref<2x3xf32> to memref<*xf32>
// CHECK: %1 = memref_cast %arg2 : memref<2x3xf32> to memref<*xf32>
// CHECK: call @__mlir_callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_softmax(%arg0: !ng.tensor<2x3xf32>, %arg1: !ng.tensor<1x!ng.i64>) -> !ng.tensor<2x3xf32> {
%0 = "ng.softmax"(%arg0) {axes = [0]} : (!ng.tensor<2x3xf32>) -> !ng.tensor<2x3xf32>
"ng.return"(%0) : (!ng.tensor<2x3xf32>) -> ()
......@@ -26,7 +26,7 @@ func @simple_softmax(%arg0: !ng.tensor<2x3xf32>, %arg1: !ng.tensor<1x!ng.i64>) -
// CHECK: %1 = memref_cast %arg1 : memref<6x4xf32> to memref<*xf32>
// CHECK: %2 = memref_cast %arg2 : memref<3x4xf32> to memref<*xf32>
// CHECK: %3 = memref_cast %arg3 : memref<3x4xf32> to memref<*xf32>
// CHECK: call @__mlir_callback_3_inputs(%0, %1, %2, %3, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_3_inputs(%0, %1, %2, %3, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_gemm(%arg0: !ng.tensor<3x6xf32>, %arg1: !ng.tensor<6x4xf32>, %arg2: !ng.tensor<3x4xf32>) -> !ng.tensor<3x4xf32> {
%0 = "ng.gemm"(%arg0, %arg1, %arg2) {alpha = 1.000000e+00 : f32, beta = 1.000000e+00 : f32, transA = false, transB = false} : (!ng.tensor<3x6xf32>, !ng.tensor<6x4xf32>, !ng.tensor<3x4xf32>) -> !ng.tensor<3x4xf32>
"ng.return"(%0) : (!ng.tensor<3x4xf32>) -> ()
......@@ -41,7 +41,7 @@ func @simple_gemm(%arg0: !ng.tensor<3x6xf32>, %arg1: !ng.tensor<6x4xf32>, %arg2:
// CHECK: %0 = memref_cast %arg0 : memref<3x2xf32> to memref<*xf32>
// CHECK: %1 = memref_cast %arg1 : memref<2x3xf32> to memref<*xf32>
// CHECK: %2 = memref_cast %arg2 : memref<2x2xf32> to memref<*xf32>
// CHECK: call @__mlir_callback_2_inputs(%0, %1, %2, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_2_inputs(%0, %1, %2, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_matmul(%arg0: !ng.tensor<3x2xf32>, %arg1: !ng.tensor<2x3xf32>) -> !ng.tensor<2x2xf32> {
%0 = "ng.matmul"(%arg0, %arg1) {transposeA = true, transposeB = true} : (!ng.tensor<3x2xf32>, !ng.tensor<2x3xf32>) -> !ng.tensor<2x2xf32>
"ng.return"(%0) : (!ng.tensor<2x2xf32>) -> ()
......@@ -55,7 +55,7 @@ func @simple_matmul(%arg0: !ng.tensor<3x2xf32>, %arg1: !ng.tensor<2x3xf32>) -> !
// CHECK: %1 = memref_cast %arg1 : memref<2x1x3x3xf32> to memref<*xf32>
// CHECK: %[[C1:.*]] = constant 0 : i64
// CHECK: %[[C2:.*]] = constant {{[0-9]+}} : i64
// CHECK: call @__mlir_callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_avgpool(%arg0: !ng.tensor<2x1x3x3xf32>) -> !ng.tensor<2x1x3x3xf32> {
%0 = "ng.avgPool"(%arg0) {includePadding = true, padAbove = [1, 1], padBelow = [0, 0], windowMovementStrides = [1, 1], windowShape = [2, 2]} : (!ng.tensor<2x1x3x3xf32>) -> !ng.tensor<2x1x3x3xf32>
"ng.return"(%0) : (!ng.tensor<2x1x3x3xf32>) -> ()
......@@ -69,7 +69,7 @@ func @simple_avgpool(%arg0: !ng.tensor<2x1x3x3xf32>) -> !ng.tensor<2x1x3x3xf32>
// CHECK: %1 = memref_cast %arg1 : memref<2x2x3x3xf32> to memref<*xf32>
// CHECK: %[[C1:.*]] = constant 0 : i64
// CHECK: %[[C2:.*]] = constant {{[0-9]+}} : i64
// CHECK: call @__mlir_callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_avgpoolbackprop(%arg0: !ng.tensor<2x2x2x2xf32>) -> !ng.tensor<2x2x3x3xf32> {
%0 = "ng.avgPoolBackprop"(%arg0) {forwardArgShape = [2, 2, 3, 3], includePadding = false, padAbove = [0, 0], padBelow = [0, 0], windowMovementStrides = [1, 1], windowShape = [2, 2]} : (!ng.tensor<2x2x2x2xf32>) -> !ng.tensor<2x2x3x3xf32>
"ng.return"(%0) : (!ng.tensor<2x2x3x3xf32>) -> ()
......@@ -83,7 +83,7 @@ func @simple_avgpoolbackprop(%arg0: !ng.tensor<2x2x2x2xf32>) -> !ng.tensor<2x2x3
// CHECK: %1 = memref_cast %arg1 : memref<64x3x9x6x5xf32> to memref<*xf32>
// CHECK: %[[C1:.*]] = constant 0 : i64
// CHECK: %[[C2:.*]] = constant {{[0-9]+}} : i64
// CHECK: call @__mlir_callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_1_input(%0, %1, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_maxpool(%arg0: !ng.tensor<64x3x7x8x10xf32>) -> !ng.tensor<64x3x9x6x5xf32> {
%0 = "ng.maxPool"(%arg0) {padAbove = [6, 4, 5], padBelow = [5, 6, 4], windowMovementStrides = [2, 3, 4], windowShape = [2, 3, 2]} : (!ng.tensor<64x3x7x8x10xf32>) -> !ng.tensor<64x3x9x6x5xf32>
"ng.return"(%0) : (!ng.tensor<64x3x9x6x5xf32>) -> ()
......@@ -98,7 +98,7 @@ func @simple_maxpool(%arg0: !ng.tensor<64x3x7x8x10xf32>) -> !ng.tensor<64x3x9x6x
// CHECK: %2 = memref_cast %arg2 : memref<2x2x5x5xf32> to memref<*xf32>
// CHECK: %[[C1:.*]] = constant 0 : i64
// CHECK: %[[C2:.*]] = constant {{[0-9]+}} : i64
// CHECK: call @__mlir_callback_2_inputs(%0, %1, %2, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, memref<*xf32>, i64, i64) -> ()
// CHECK: call @callback_2_inputs(%0, %1, %2, %[[C1]], %[[C2]]) : (memref<*xf32>, memref<*xf32>, memref<*xf32>, i64, i64) -> ()
func @simple_maxpoolbackprop(%arg0: !ng.tensor<2x2x5x5xf32>, %arg1: !ng.tensor<2x2x4x3xf32>) -> !ng.tensor<2x2x5x5xf32> {
%0 = "ng.maxPoolBackprop"(%arg0, %arg1) {padAbove = [0, 0], padBelow = [0, 0], windowMovementStrides = [1, 1], windowShape = [2, 3]} : (!ng.tensor<2x2x5x5xf32>, !ng.tensor<2x2x4x3xf32>) -> !ng.tensor<2x2x5x5xf32>
"ng.return"(%0) : (!ng.tensor<2x2x5x5xf32>) -> ()
......
test/models/onnx/dynamic_shapes/average_pool_2d_dyn.prototxt 0 → 100644
View file @ 310fcf07
ir_version: 3
producer_name: "nGraph ONNX Importer"
graph {
node {
input: "x"
output: "y"
op_type: "AveragePool"
attribute {
name: "kernel_shape"
ints: 2
ints: 2
type: INTS
}
attribute {
name: "strides"
ints: 2
ints: 2
type: INTS
}
}
name: "compute_graph"
input {
name: "x"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
}
}
}
}
output {
name: "y"
type {
tensor_type {
elem_type: 1
shape {
}
}
}
}
}
opset_import {
version: 7
}
test/models/onnx/dynamic_shapes/global_average_pool_dyn.prototxt 0 → 100644
View file @ 310fcf07
ir_version: 3
producer_name: "nGraph ONNX Importer"
graph {
node {
input: "x"
output: "y"
op_type: "GlobalAveragePool"
attribute {
name: "strides"
ints: 2
ints: 2
type: INTS
}
}
name: "compute_graph"
input {
name: "x"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
dim {
dim_value: 5
}
dim {
dim_value: 5
}
}
}
}
}
output {
name: "y"
type {
tensor_type {
elem_type: 1
shape {
}
}
}
}
}
opset_import {
version: 7
}
test/models/onnx/dynamic_shapes/global_max_pool_dyn.prototxt 0 → 100644
View file @ 310fcf07
ir_version: 3
producer_name: "nGraph ONNX Importer"
graph {
node {
input: "x"
output: "y"
op_type: "GlobalMaxPool"
attribute {
name: "strides"
ints: 2
ints: 2
type: INTS
}
}
name: "compute_graph"
input {
name: "x"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
dim {
dim_value: 5
}
dim {
dim_value: 5
}
}
}
}
}
output {
name: "y"
type {
tensor_type {
elem_type: 1
shape {
}
}
}
}
}
opset_import {
version: 7
}
test/models/onnx/dynamic_shapes/max_pool_2d_dyn.prototxt 0 → 100644
View file @ 310fcf07
ir_version: 3
producer_name: "nGraph ONNX Importer"
graph {
node {
input: "x"
output: "y"
op_type: "MaxPool"
attribute {
name: "kernel_shape"
ints: 2
ints: 2
type: INTS
}
attribute {
name: "strides"
ints: 2
ints: 2
type: INTS
}
attribute {
name: "pads"
ints: 1
ints: 1
ints: 1
ints: 1
type: INTS
}
}
name: "compute_graph"
input {
name: "x"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
dim {
dim_param: "batch"
}
}
}
}
}
output {
name: "y"
type {
tensor_type {
elem_type: 1
shape {
}
}
}
}
}
opset_import {
version: 7
}
test/models/onnx/round.prototxt 0 → 100644
View file @ 310fcf07
ir_version: 3
producer_name: "backend-test"
graph {
node {
input: "x"
output: "y"
op_type: "Round"
}
name: "test_round"
input {
name: "x"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_value: 15
}
}
}
}
}
output {
name: "y"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_value: 15
}
}
}
}
}
}
opset_import {
version: 11
}
test/onnx/onnx_import.in.cpp
View file @ 310fcf07
......@@ -1963,3 +1963,30 @@ NGRAPH_TEST(onnx_${BACKEND_NAME}, model_reciprocal)
test_case.run();
}
NGRAPH_TEST(onnx_${BACKEND_NAME}, model_round)
{
const auto round_fn =
onnx_import::import_onnx_model(file_util::path_join(SERIALIZED_ZOO, "onnx/round.prototxt"));
auto test_case = ngraph::test::NgraphTestCase(round_fn, "${BACKEND_NAME}");
test_case.add_input<float>({0.1f,
0.5f,
0.9f,
1.2f,
1.5f,
1.8f,
2.3f,
2.5f,
2.7f,
-1.1f,
-1.5f,
-1.9f,
-2.2f,
-2.5f,
-2.8f});
test_case.add_expected_output<float>(
{0.f, 0.f, 1.f, 1.f, 2.f, 2.f, 2.f, 2.f, 3.f, -1.f, -2.f, -2.f, -2.f, -2.f, -3.f});
test_case.run();
}
test/onnx/onnx_import_dyn_shapes.in.cpp
View file @ 310fcf07
......@@ -282,3 +282,83 @@ NGRAPH_TEST(onnx_dyn_shapes_${BACKEND_NAME}, model_conv_with_dynamic_batch)
test_case.run();
}
NGRAPH_TEST(onnx_dyn_shapes_${BACKEND_NAME}, avg_pool_dyn_shape)
{
const auto function = onnx_import::import_onnx_model(
file_util::path_join(SERIALIZED_ZOO, "onnx/dynamic_shapes/average_pool_2d_dyn.prototxt"));
auto test_case = NgraphTestCase(function, "${BACKEND_NAME}", BackendMode::DYNAMIC);
const Shape shape{1, 1, 4, 4};
const auto elems_in_tensor = shape_size(shape);
std::vector<float> input_values(elems_in_tensor);
std::iota(input_values.begin(), input_values.end(), 0.f);
test_case.add_input<float>(shape, input_values);
std::vector<float> expected_values{2.5f, 4.5f, 10.5f, 12.5f};
test_case.add_expected_output<float>(Shape{1, 1, 2, 2}, expected_values);
test_case.run();
}
NGRAPH_TEST(onnx_dyn_shapes_${BACKEND_NAME}, max_pool_dyn_shape)
{
const auto function = onnx_import::import_onnx_model(
file_util::path_join(SERIALIZED_ZOO, "onnx/dynamic_shapes/max_pool_2d_dyn.prototxt"));
auto test_case = NgraphTestCase(function, "${BACKEND_NAME}", BackendMode::DYNAMIC);
const Shape shape{1, 1, 4, 4};
const auto elems_in_tensor = shape_size(shape);
std::vector<float> input_values(elems_in_tensor);
std::iota(input_values.begin(), input_values.end(), 0.f);
test_case.add_input<float>(shape, input_values);
std::vector<float> expected_values{0.f, 2.f, 3.f, 8.f, 10.f, 11.f, 12.f, 14.f, 15.f};
test_case.add_expected_output<float>(Shape{1, 1, 3, 3}, expected_values);
test_case.run();
}
NGRAPH_TEST(onnx_dyn_shapes_${BACKEND_NAME}, global_avg_pool_dyn_shape)
{
const auto function = onnx_import::import_onnx_model(file_util::path_join(
SERIALIZED_ZOO, "onnx/dynamic_shapes/global_average_pool_dyn.prototxt"));
auto test_case = NgraphTestCase(function, "${BACKEND_NAME}", BackendMode::DYNAMIC);
const Shape shape{1, 3, 5, 5};
const auto elems_in_tensor = shape_size(shape);
std::vector<float> input_values(elems_in_tensor);
std::iota(input_values.begin(), input_values.end(), 0.f);
test_case.add_input<float>(shape, input_values);
std::vector<float> expected_values{12.f, 37.f, 62.f};
test_case.add_expected_output<float>(Shape{1, 3, 1, 1}, expected_values);
test_case.run();
}
NGRAPH_TEST(onnx_dyn_shapes_${BACKEND_NAME}, global_max_pool_dyn_shape)
{
const auto function = onnx_import::import_onnx_model(
file_util::path_join(SERIALIZED_ZOO, "onnx/dynamic_shapes/global_max_pool_dyn.prototxt"));
auto test_case = NgraphTestCase(function, "${BACKEND_NAME}", BackendMode::DYNAMIC);
const Shape shape{1, 3, 5, 5};
const auto elems_in_tensor = shape_size(shape);
std::vector<float> input_values(elems_in_tensor);
std::iota(input_values.begin(), input_values.end(), 0.f);
test_case.add_input<float>(shape, input_values);
std::vector<float> expected_values{24.f, 49.f, 74.f};
test_case.add_expected_output<float>(Shape{1, 3, 1, 1}, expected_values);
test_case.run();
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment