Problem when implementing trained TensorFlow Model in Simulink using the TensorFlow Model Predict block.
3 views (last 30 days)
Show older comments
I have previously trained a TensorFlow model in python and I'm looking to implement it into Simulink, which I thought was supposed to be straight forward. However I ran into the following error where it says the output from the predict block is producing variable size signal (they are supposed to be all 1 by 1 float 32)

I don't understand what the problem may be. I have set the maximum Matlab dim size to 1s as the following.


Any Ideas what the problem might be?
1 Comment
Gayathri
on 24 Dec 2024
Hi @Bevan Bintang, can you please provide the model and the trained model file so that I can look into the issue in detail
Answers (1)
aditi bagora
on 6 Jan 2025
It seems there might be a configuration issue between the 'Output_Port_1' of the bus selector and the 'Input_port_1' of the 'Moment_aft' subsystem. This could be due to the 'Tensorflow Model Predict' block generating a variable-size output, which the subsystem might not support or be configured to handle.
As @Gayathri suggested, to help resolve the issue, it would be beneficial if you could provide the necessary files.
Hope this helps!
0 Comments
See Also
Categories
Find more on Deep Learning Toolbox in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!