TMVA::Experimental::SOFIE::PyTorch::INTERNAL Namespace Reference

Typedefs
using	PyTorchMethodMap = std::unordered_map< std::string, std::unique_ptr< ROperator >(*)(PyObject *fNode)>

Functions
std::unique_ptr< ROperator >	MakePyTorchConv (PyObject *fNode)
	Prepares a ROperator_Conv object.
std::unique_ptr< ROperator >	MakePyTorchGemm (PyObject *fNode)
	Prepares a ROperator_Gemm object.
std::unique_ptr< ROperator >	MakePyTorchNode (PyObject *fNode)
	Prepares equivalent ROperator with respect to PyTorch ONNX node.
std::unique_ptr< ROperator >	MakePyTorchRelu (PyObject *fNode)
	Prepares a ROperator_Relu object.
std::unique_ptr< ROperator >	MakePyTorchSelu (PyObject *fNode)
	Prepares a ROperator_Selu object.
std::unique_ptr< ROperator >	MakePyTorchSigmoid (PyObject *fNode)
	Prepares a ROperator_Sigmoid object.
std::unique_ptr< ROperator >	MakePyTorchTranspose (PyObject *fNode)
	Prepares a ROperator_Transpose object.

Variables
const PyTorchMethodMap	mapPyTorchNode

Typedef Documentation

PyTorchMethodMap

using TMVA::Experimental::SOFIE::PyTorch::INTERNAL::PyTorchMethodMap = typedef std::unordered_map<std::string, std::unique_ptr<ROperator> (*)(PyObject* fNode)>

Definition at line 55 of file RModelParser_PyTorch.cxx.

Function Documentation

MakePyTorchConv()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchConv

(

PyObject *

fNode

)

Prepares a ROperator_Conv object.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

Unique pointer to ROperator object

For Conv Operator of PyTorch's ONNX Graph, attributes like dilations, group, kernel shape, pads and strides are found, and are passed in instantiating the ROperator object with autopad default to NOTSET.

Definition at line 276 of file RModelParser_PyTorch.cxx.

MakePyTorchGemm()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchGemm

(

PyObject *

fNode

)

Prepares a ROperator_Gemm object.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

Unique pointer to ROperator object

For PyTorch's Linear layer having Gemm operation in its ONNX graph, the names of the input tensor, output tensor are extracted, and then are passed to instantiate a ROperator_Gemm object using the required attributes. fInputs is a list of tensor names, which includes the names of the input tensor and the weight tensors.

Definition at line 114 of file RModelParser_PyTorch.cxx.

MakePyTorchNode()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchNode

(

PyObject *

fNode

)

Prepares equivalent ROperator with respect to PyTorch ONNX node.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

unique pointer to ROperator object

Function searches for the passed PyTorch ONNX Graph node in the map, and calls the specific preparatory function, subsequently returning the ROperator object.

For developing new preparatory functions for supporting PyTorch ONNX Graph nodes in future, all one needs is to extract the required properties and attributes from the fNode dictionary which contains all the information about any PyTorch ONNX the ROperator object.

The fNode dictionary which holds all the information about a PyTorch ONNX Graph's node has following structure:-

dict fNode {  'nodeType'        : Type of node (operator)
              'nodeAttributes'  : Attributes of the node
              'nodeInputs'      : List of names of input tensors
              'nodeOutputs'     : List of names of output tensors
              'nodeDType'       : Data-type of the operator node
           }

Definition at line 93 of file RModelParser_PyTorch.cxx.

MakePyTorchRelu()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchRelu

(

PyObject *

fNode

)

Prepares a ROperator_Relu object.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

Unique pointer to ROperator object

For instantiating a ROperator_Relu object, the names of input & output tensors and the data-type of the Graph node are extracted.

Definition at line 160 of file RModelParser_PyTorch.cxx.

MakePyTorchSelu()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchSelu

(

PyObject *

fNode

)

Prepares a ROperator_Selu object.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

Unique pointer to ROperator object

For instantiating a ROperator_Selu object, the names of input & output tensors and the data-type of the Graph node are extracted.

Definition at line 187 of file RModelParser_PyTorch.cxx.

MakePyTorchSigmoid()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchSigmoid

(

PyObject *

fNode

)

Prepares a ROperator_Sigmoid object.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

Unique pointer to ROperator object

For instantiating a ROperator_Sigmoid object, the names of input & output tensors and the data-type of the Graph node are extracted.

Definition at line 213 of file RModelParser_PyTorch.cxx.

MakePyTorchTranspose()

std::unique_ptr< ROperator > TMVA::Experimental::SOFIE::PyTorch::INTERNAL::MakePyTorchTranspose

(

PyObject *

fNode

)

Prepares a ROperator_Transpose object.

Parameters

[in]

fNode

Python PyTorch ONNX Graph node

Returns

Unique pointer to ROperator object

For Transpose Operator of PyTorch's ONNX Graph, the permute dimensions are found, and are passed in instantiating the ROperator object.

Definition at line 239 of file RModelParser_PyTorch.cxx.

Variable Documentation

mapPyTorchNode

const PyTorchMethodMap TMVA::Experimental::SOFIE::PyTorch::INTERNAL::mapPyTorchNode

Initial value:

=
{
    {"onnx::Gemm",      &MakePyTorchGemm},
    {"onnx::Conv",      &MakePyTorchConv},
    {"onnx::Relu",      &MakePyTorchRelu},
    {"onnx::Selu",      &MakePyTorchSelu},
    {"onnx::Sigmoid",   &MakePyTorchSigmoid},
    {"onnx::Transpose", &MakePyTorchTranspose}
}

Definition at line 57 of file RModelParser_PyTorch.cxx.