Current Problem
Implement ReduceSum mathematical Operation
Rules
ONNX standard
Every math method must respect the onnx representation.
You can find all the necessary details about onnx operators here !
Standard & lean format
The "standard version" of a method is in the format:
pub fn methodName (args) !Tensor(T) {
...
outputShape = compute output shape...
outputTensor = compute output tensor...
...
methodName_lean(input, args..., output)
return outputTensor;
}
It is suggested to write a get_methodName_oputput_shape() to compute the shape of the output so that the same method can be called during the generation of the readyGraph.
The "lean version" of a method is in the format:
pub inline fn methodName_lean (args) void {
...
... actual computation of the output
}
"lean" and "standard" are mandatory.
The lean version is called during the codegen and the standard version is used in testings.
Current Problem
Implement ReduceSum mathematical Operation
Rules
ONNX standard
Every math method must respect the onnx representation.
You can find all the necessary details about onnx operators here !
Standard & lean format
The "standard version" of a method is in the format:
It is suggested to write a
get_methodName_oputput_shape()to compute the shape of the output so that the same method can be called during the generation of the readyGraph.The "lean version" of a method is in the format:
"lean" and "standard" are mandatory.
The lean version is called during the codegen and the standard version is used in testings.