[mlir][tosa] Replace UniformQuantizedType by the more generic Quantiz… #126275
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-mlir-tosa Author: Tai Ly (Tai78641) Changes…edType in Conv verifiers also fixed buildTransConvOpWithQuantInfo to insert input/weight zp operands Change-Id: Ie1961af931864f801914a62976bc988881ee075e Full diff: https://github.com/llvm/llvm-project/pull/126275.diff 1 Files Affected:
diff --git a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
index 031c279ff09e275..6143a9d23a00394 100644
--- a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
@@ -240,16 +240,13 @@ static LogicalResult verifyConvOp(T op) {
bool biasIsFloat = llvm::isa<FloatType>(biasEType);
bool resultIsFloat = llvm::isa<FloatType>(resultEType);
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(inputEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(inputEType))
inputEType = quantType.getStorageType();
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(biasEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(biasEType))
biasEType = quantType.getStorageType();
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(resultEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(resultEType))
resultEType = quantType.getStorageType();
if (biasIsFloat && resultIsFloat && (biasEType != resultEType)) {
@@ -346,8 +343,7 @@ static LogicalResult verifyConvOpModes(T op) {
auto inputEType =
llvm::cast<ShapedType>(op.getInput().getType()).getElementType();
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(inputEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(inputEType))
inputEType = quantType.getStorageType();
auto accType = op.getAccType();
@@ -369,7 +365,23 @@ static LogicalResult verifyConvOpModes(T op) {
if (inputEType.isF32() && !accType.isF32())
return op.emitOpError("accumulator type for f32 tensor is not f32");
- return success();
+ auto resultEType =
+ llvm::cast<ShapedType>(op.getResult().getType()).getElementType();
+
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(resultEType))
+ resultEType = quantType.getStorageType();
+
+ // check allowed input/result element types combinations
+ if ((inputEType.isInteger(8) && resultEType.isInteger(32)) ||
+ (inputEType.isInteger(16) && resultEType.isInteger(48)) ||
+ (isa<Float8E5M2Type>(inputEType) && resultEType.isF16()) ||
+ (isa<Float8E4M3FNType>(inputEType) && resultEType.isF16()) ||
+ (inputEType.isF16() && resultEType.isF16()) ||
+ (inputEType.isBF16() && resultEType.isBF16()) ||
+ (inputEType.isF32() && resultEType.isF32()))
+ return success();
+
+ return op.emitOpError("input/output element types are incompatible.");
}
// verify that inType and outType have same element types
@@ -519,7 +531,8 @@ static void buildTransConvOpWithQuantInfo(
OpBuilder &builder, OperationState &result, Type outputType, Value input,
Value weight, Value bias, DenseI64ArrayAttr outpad,
DenseI64ArrayAttr stride, DenseI64ArrayAttr outputShape, TypeAttr accType) {
- result.addOperands({input, weight, bias});
+ auto zps = createZPsAsConst(builder, input, weight);
+ result.addOperands({input, weight, bias, zps.first, zps.second});
result.addAttribute("out_pad", outpad);
result.addAttribute("stride", stride);
result.addAttribute("out_shape", outputShape);
@@ -2478,18 +2491,15 @@ LogicalResult mlir::tosa::getZeroPoint(ElementsAttr zpAttr, int64_t &zp) {
return failure();
}
-// Create a rank-0 const tensor for zero point of the source tensor.
+// Create a rank-1 const tensor for zero point of the source tensor.
std::optional<Value> mlir::tosa::createZeroPointTensor(OpBuilder &builder,
Location loc,
Type srcElemType,
int64_t zp) {
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(srcElemType))
- srcElemType = quantType.getStorageType();
-
- auto zpType = mlir::RankedTensorType::get({1}, srcElemType);
+ srcElemType = getElementTypeOrSelf(srcElemType);
if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(srcElemType))
srcElemType = quantType.getStorageType();
+ auto zpType = mlir::RankedTensorType::get({1}, srcElemType);
if (llvm::isa<FloatType>(srcElemType)) {
auto zpAttr = DenseElementsAttr::get(
zpType, builder.getFloatAttr(srcElemType, static_cast<double>(zp)));
|
|
@llvm/pr-subscribers-mlir Author: Tai Ly (Tai78641) Changes…edType in Conv verifiers also fixed buildTransConvOpWithQuantInfo to insert input/weight zp operands Change-Id: Ie1961af931864f801914a62976bc988881ee075e Full diff: https://github.com/llvm/llvm-project/pull/126275.diff 1 Files Affected:
diff --git a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
index 031c279ff09e275..6143a9d23a00394 100644
--- a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
@@ -240,16 +240,13 @@ static LogicalResult verifyConvOp(T op) {
bool biasIsFloat = llvm::isa<FloatType>(biasEType);
bool resultIsFloat = llvm::isa<FloatType>(resultEType);
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(inputEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(inputEType))
inputEType = quantType.getStorageType();
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(biasEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(biasEType))
biasEType = quantType.getStorageType();
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(resultEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(resultEType))
resultEType = quantType.getStorageType();
if (biasIsFloat && resultIsFloat && (biasEType != resultEType)) {
@@ -346,8 +343,7 @@ static LogicalResult verifyConvOpModes(T op) {
auto inputEType =
llvm::cast<ShapedType>(op.getInput().getType()).getElementType();
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(inputEType))
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(inputEType))
inputEType = quantType.getStorageType();
auto accType = op.getAccType();
@@ -369,7 +365,23 @@ static LogicalResult verifyConvOpModes(T op) {
if (inputEType.isF32() && !accType.isF32())
return op.emitOpError("accumulator type for f32 tensor is not f32");
- return success();
+ auto resultEType =
+ llvm::cast<ShapedType>(op.getResult().getType()).getElementType();
+
+ if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(resultEType))
+ resultEType = quantType.getStorageType();
+
+ // check allowed input/result element types combinations
+ if ((inputEType.isInteger(8) && resultEType.isInteger(32)) ||
+ (inputEType.isInteger(16) && resultEType.isInteger(48)) ||
+ (isa<Float8E5M2Type>(inputEType) && resultEType.isF16()) ||
+ (isa<Float8E4M3FNType>(inputEType) && resultEType.isF16()) ||
+ (inputEType.isF16() && resultEType.isF16()) ||
+ (inputEType.isBF16() && resultEType.isBF16()) ||
+ (inputEType.isF32() && resultEType.isF32()))
+ return success();
+
+ return op.emitOpError("input/output element types are incompatible.");
}
// verify that inType and outType have same element types
@@ -519,7 +531,8 @@ static void buildTransConvOpWithQuantInfo(
OpBuilder &builder, OperationState &result, Type outputType, Value input,
Value weight, Value bias, DenseI64ArrayAttr outpad,
DenseI64ArrayAttr stride, DenseI64ArrayAttr outputShape, TypeAttr accType) {
- result.addOperands({input, weight, bias});
+ auto zps = createZPsAsConst(builder, input, weight);
+ result.addOperands({input, weight, bias, zps.first, zps.second});
result.addAttribute("out_pad", outpad);
result.addAttribute("stride", stride);
result.addAttribute("out_shape", outputShape);
@@ -2478,18 +2491,15 @@ LogicalResult mlir::tosa::getZeroPoint(ElementsAttr zpAttr, int64_t &zp) {
return failure();
}
-// Create a rank-0 const tensor for zero point of the source tensor.
+// Create a rank-1 const tensor for zero point of the source tensor.
std::optional<Value> mlir::tosa::createZeroPointTensor(OpBuilder &builder,
Location loc,
Type srcElemType,
int64_t zp) {
- if (auto quantType =
- llvm::dyn_cast<mlir::quant::UniformQuantizedType>(srcElemType))
- srcElemType = quantType.getStorageType();
-
- auto zpType = mlir::RankedTensorType::get({1}, srcElemType);
+ srcElemType = getElementTypeOrSelf(srcElemType);
if (auto quantType = llvm::dyn_cast<mlir::quant::QuantizedType>(srcElemType))
srcElemType = quantType.getStorageType();
+ auto zpType = mlir::RankedTensorType::get({1}, srcElemType);
if (llvm::isa<FloatType>(srcElemType)) {
auto zpAttr = DenseElementsAttr::get(
zpType, builder.getFloatAttr(srcElemType, static_cast<double>(zp)));
|
Tai78641
force-pushed
the
pr_use_quant
branch
from
b67944b to
c62371b
Compare
GeorgeARM
requested changes
Feb 7, 2025
…edType in Conv verifiers Change-Id: Ie1961af931864f801914a62976bc988881ee075e Signed-off-by: Tai Ly <[email protected]>
Tai78641
force-pushed
the
pr_use_quant
branch
from
c62371b to
dd71961
Compare
added tests |
GeorgeARM
approved these changes
Feb 11, 2025
Icohedron
pushed a commit
to Icohedron/llvm-project
that referenced
this pull request
Feb 11, 2025
Replace UniformQuantizedType by the more generic QuantizedType in Conv verifiers. Change-Id: Ie1961af931864f801914a62976bc988881ee075e Signed-off-by: Tai Ly <[email protected]> Co-authored-by: Thibaut Goetghebuer-Planchon <[email protected]>
joaosaffran
pushed a commit
to joaosaffran/llvm-project
that referenced
this pull request
Feb 14, 2025
Replace UniformQuantizedType by the more generic QuantizedType in Conv verifiers. Change-Id: Ie1961af931864f801914a62976bc988881ee075e Signed-off-by: Tai Ly <[email protected]> Co-authored-by: Thibaut Goetghebuer-Planchon <[email protected]>
sivan-shani
pushed a commit
to sivan-shani/llvm-project
that referenced
this pull request
Feb 24, 2025
Replace UniformQuantizedType by the more generic QuantizedType in Conv verifiers. Change-Id: Ie1961af931864f801914a62976bc988881ee075e Signed-off-by: Tai Ly <[email protected]> Co-authored-by: Thibaut Goetghebuer-Planchon <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
…edType in Conv verifiers
Change-Id: Ie1961af931864f801914a62976bc988881ee075e