saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-05-24 22:17:32 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

[WebNN EP] Update data type constraints for Reduction ops (#20912)

Browse source 
WebNN Spec adds missing 64-bit integers support for `reduceL1`,
`reduceSum`, `reduceSumSquare` and `reduceProduct` ops at this
[PR](https://github.com/webmachinelearning/webnn/pull/695), which has
already been implemented in Chromium. Update corresponding data type
constraints in WebNN EP.

Besides, WebNN CPU backend currently doesn't support `uint64` and
`uint32` for these ops.
This commit is contained in:
Wanming Lin 2024-06-17 23:46:18 +08:00 committed by GitHub
parent 8aa2667ae6
commit bbb6dbf6d2
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 7 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
onnxruntime/core/providers/webnn/builders/impl/reduction_op_builder.cc
View file

@ -162,10 +162,15 @@ bool ReductionOpBuilder::HasSupportedInputsImpl(const Node& node, const WebnnDev
ONNX_NAMESPACE::TensorProto_DataType_FLOAT16,
ONNX_NAMESPACE::TensorProto_DataType_INT32,
ONNX_NAMESPACE::TensorProto_DataType_UINT32,
ONNX_NAMESPACE::TensorProto_DataType_INT64,
ONNX_NAMESPACE::TensorProto_DataType_UINT64,
};
// WebNN CPU backend doesn't support uint32 for reduceProd and reduceSum.
if (device_type == WebnnDeviceType::CPU && (op_type == "ReduceProd" || op_type == "ReduceSum")) {
if (device_type == WebnnDeviceType::CPU) {
// WebNN CPU backend doesn't support uint32 and uint64 for reduceL1,
// reduceProd, reduceSum and reduceSumSquare.
supported_data_types.erase(ONNX_NAMESPACE::TensorProto_DataType_UINT32);
supported_data_types.erase(ONNX_NAMESPACE::TensorProto_DataType_UINT64);
}
} else if (op_type == "ReduceL2" || op_type == "ReduceLogSum" ||
op_type == "ReduceLogSumExp" || op_type == "ReduceMean") {