Is it possible to add fine-grained operator export operations to the "exportONNXNetwork" function in a future release?

Is it possible to add fine-grained operator export operations to the "exportONNXNetwork" function in a future release?

在未来版本中"exportONNXNetwork"函数是否可以增加对细粒度的operator导出操作？

This is not possible in the latest version R2021a

For example, the following simple example would be supported:

a  = dlarray(1:10);
b = dlarray(2:11);
c = a+b;
exportONNXNetwork(@(x,y)mysum,a,b)% unsupported???

Error using nnet.internal.cnn.onnx.exportONNXNetwork>iValidateInputs (line 49)
Expected a string scalar or character vector for the parameter name.

Error in nnet.internal.cnn.onnx.exportONNXNetwork (line 29)
[NNTNetwork, Filename, NetworkName, OpsetVersion] = iValidateInputs(NNTNetwork, Filename, defaultOpset, varargin{:});

Error in exportONNXNetwork (line 38)
nnet.internal.cnn.onnx.exportONNXNetwork(Network, filename, varargin{:});

function c = mysum(a,b)
c = a+b;
end

Answer 1

[cui,xingxing]

It is believed that support for this feature can be added in future versions

Answer 2

[Steven Lord]

Generally we don't comment on whether something requested is planned or being implemented for a future release. [In this case, I honestly don't know.] If you believe this functionality should be added to Deep Learning Toolbox, please file an enhancement request with Technical Support using the Contact Support link on the Support section of this website.