Unable to specify BatchSize in exportONNXNetwork

3 次查看（过去 30 天）

Gabor Balazs 2022-3-22

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1677389-unable-to-specify-batchsize-in-exportonnxnetwork

评论： Sivylla Paraskevopoulou 2022-5-9

Hello!

I want to export a simple neural network from Matlab to the .ONNX format. Therefore, I use the exportONNXNetwork function. In the documentation, there is the name-value argument "BatchSize", which allows to fix the batch size of the network, which I need to be set to 1.

exportONNXNetwork(net, 'network.onnx', "BatchSize", 1)

Sadly, I cannot set the name-value argument as described in the documentation (https://www.mathworks.com/help/deeplearning/ref/exportonnxnetwork.html); it is not defined inside the exportONNXNetwork function anywhere and finally throws an error in iValidateInputs:

Error using nnet.internal.cnn.onnx.exportONNXNetwork>iValidateInputs (line 49)
'BatchSize' is not a recognized parameter. For a list of valid name-value pair arguments, see the documentation for this function.
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{:});
Error in eval_mlp (line 76)
exportONNXNetwork(net, 'network.onnx', "BatchSize", 1)