我从pytorch中​导入net到工作区，​在把net导入到si​mulink中的ma​tlab function中，​在运行过程中出现如下​问题Code generation for custom layer 'aten__linear0' for target 'mkldnn' is not supported as it returns a dlarra

classdef aten__linear0 < nnet.layer.Layer & nnet.layer.Formattable & ...

nnet.layer.AutogeneratedFromPyTorch & nnet.layer.Acceleratable

%aten__linear0 Auto-generated custom layer

% Auto-generated by MATLAB on 2024-05-24 16:18:35

%#codegen

properties (Learnable)

% Networks (type dlnetwork)

end

properties

% Non-Trainable Parameters

end

properties (Learnable)

% Trainable Parameters

Param_weight

Param_bias

end

methods

function obj = aten__linear0(Name, Type, InputNames, OutputNames)

obj.Name = Name;

obj.Type = Type;

obj.NumInputs = 1;

obj.NumOutputs = 1;

obj.InputNames = InputNames;

obj.OutputNames = OutputNames;

end

function [linear_9] = predict(obj,linear_x_1)

%Validates that the input has the correct format and permutes its dimensions into the reverse of the original PyTorch format.

model_tt.ops.validateInput(linear_x_1,2);

[linear_x_1, linear_x_1_format] = model_tt.ops.permuteInputToReversePyTorch(linear_x_1, 2);

[linear_x_1] = struct('value', linear_x_1, 'rank', int64(2));

[linear_9] = tracedPyTorchFunction(obj,linear_x_1,false,"predict");

%Permute U-labelled output to forward PyTorch dimension ordering

if(any(dims(linear_9.value) == 'U'))

linear_9 = permute(linear_9.value, fliplr(1:max(2,linear_9.rank)));

end

end

function [linear_9] = forward(obj,linear_x_1)

%Validates that the input has the correct format and permutes its dimensions into the reverse of the original PyTorch format.

model_tt.ops.validateInput(linear_x_1,2);

[linear_x_1, linear_x_1_format] = model_tt.ops.permuteInputToReversePyTorch(linear_x_1, 2);

[linear_x_1] = struct('value', linear_x_1, 'rank', int64(2));

[linear_9] = tracedPyTorchFunction(obj,linear_x_1,true,"forward");

%Permute U-labelled output to forward PyTorch dimension ordering

if(any(dims(linear_9.value) == 'U'))

linear_9 = permute(linear_9.value, fliplr(1:max(2,linear_9.rank)));

end

end

function [linear_9] = tracedPyTorchFunction(obj,linear_x_1,isForward,predict)

linear_weight_1 = obj.Param_weight;

[linear_weight_1] = struct('value', dlarray(linear_weight_1,'UU'), 'rank', 2);

linear_bias_1 = obj.Param_bias;

[linear_bias_1] = struct('value', dlarray(linear_bias_1,'UU'), 'rank', 1);

[linear_9] = model_tt.ops.pyLinear(linear_x_1, linear_weight_1, linear_bias_1);

end

end

end

Answer 1

[Shivani]

Hi @guiyang,

The error message you're seeing indicates that the custom layer 'aten__linear0', which corresponds to a linear (fully connected) layer in PyTorch, is not supported for code generation for the target 'mkldnn'. MKLDNN is a backend optimized for deep learning operations on Intel CPUs, and it seems the issue is with generating code that can leverage this optimization for the specified layer.

From my understanding, a custom layer will not be supported for Code Generation by default. Please refer to the following documentation for more information on extending Code Generation support to custom layers:https://www.mathworks.com/help/coder/ug/networks-and-layers-supported-for-c-code-generation.html#:~:text=Yes-,Custom%20layers,-Custom%20layers%2C%20with

Additionally, if possible, replace the unsupported custom layer ('aten__linear0') with an equivalent operation or layer that is supported for code generation in MATLAB. The following documentation link lists out all the networks and layers supported for Code Generation: https://www.mathworks.com/help/coder/ug/networks-and-layers-supported-for-c-code-generation.html

Another possible workaround would be to generate generic C or C++ code that does not depend on third-party libraries, without targeting MKLDNN optimizations. You can refer to the following documentation link for more information on this: https://www.mathworks.com/help/coder/ug/generate-generic-cc-code-for-deep-learning-networks.html