how can ı use "minibatch​predict(ne​t,XTest);" command on simulink?

7 次查看（过去 30 天）
Bahadir
Bahadir 2025-9-29
评论： Spoorthy Kannur 2025-11-11
I trained a LSTM network.
How can I use "scores = minibatchpredict(net,XTest);" and "YPred = predict(net, XTest);" commands on Simulink?

请先登录，再回答此问题。

回答（1 个）

AJ Ibraheem
AJ Ibraheem 2025-10-6
编辑：Walter Roberson 2025-10-6
The 'Stateful Predict' block might be what you're looking for. See https://uk.mathworks.com/help/deeplearning/ref/statefulpredict.html
  5 个评论
显示 3更早的评论
Bahadir
Bahadir 2025-10-8
@AJ Ibraheem
Thank you for your answer.
could you give more detail information about how to get same result on simulink. How to use predict command at matlab function block on simulink.
function y= fnc(u)
persistent net
if isempty(net)
net = coder.loadDeepLearningNetwork('32.mat');
end
input= [u];
input=rescale(input);
XTrain = {input'};
output= predict(net, XTrain);
y=output{1};
end
Spoorthy Kannur
Spoorthy Kannur 2025-11-11
Hi Bahadir,
You may try the following:
In Simulink, you can use your trained network for prediction inside a MATLAB Function block, but there are a few important details to ensure it behaves consistently with MATLAB, in your case:
function y = fnc(u)
persistent net
if isempty(net)
net = coder.loadDeepLearningNetwork('32.mat');
end
% Preprocess input the same way as during training
input = rescale(u);
XTrain = {input'};
% Perform prediction
YPred = predict(net, XTrain);
y = YPred{1};
end
1. Use a supported compiler: “minibatchpredict” ( https://www.mathworks.com/help/deeplearning/ref/minibatchpredict.html) is not codegen-compatible, but “predict” is (https://www.mathworks.com/help/deeplearning/ref/dlnetwork.predict.html). Select a supported compiler using (Visual Studio C++ is required; MinGW64 won’t work for deep learning code generation):
mex -setup cpp
2. Match data preprocessing: Apply the same scaling or reshaping you used during training (e.g., sequence dimension order).
3. Choose the right block execution rate: For sequence data, ensure the Simulink sample time matches your network input timestep.
If your results still differ slightly from MATLAB, check whether the MATLAB version of “predict” was run statefully or statelessly, since LSTMs maintain hidden states across calls — this can cause small output differences unless you reset or manage the network state manually in Simulink.
If this does resolve the issue, kindly reach out to MathWorks Technical Support for more help (https://www.mathworks.com/support/contact_us.html)

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 的更多信息

标签

产品


版本

R2024a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by