Problems with installation of ARM Compute Library in Raspbian bookworm 32 bit

When i tried to install the library using the script installARMComputeLIB from the Harware Setup window, I receve the next error.

In file included from ./utils/Utils.h:36,

from ./utils/ImageLoader.h:32,

from utils/GraphUtils.cpp:34:

include/libnpy/npy.hpp: In function ‘std::string npy::pyparse::write_tuple(const std::vector<T>&)’:

include/libnpy/npy.hpp:298:42: error: ‘ostream_iterator’ is not a member of ‘std’

298 | std::copy(v.begin(), v.end()-1, std::ostream_iterator<T>(ss, delimiter.c_str()));

| ^~~~~~~~~~~~~~~~

include/libnpy/npy.hpp:33:1: note: ‘std::ostream_iterator’ is defined in header ‘<iterator>’; did you forget to ‘#include <iterator>’?

32 | #include <cstring>

+++ |+#include <iterator>

33 | #include <vector>

include/libnpy/npy.hpp:298:60: error: expected primary-expression before ‘>’ token

298 | std::copy(v.begin(), v.end()-1, std::ostream_iterator<T>(ss, delimiter.c_str()));

| ^

scons: *** [build/utils/GraphUtils.o] Error 1

scons: building terminated because of errors.

This error is solver in the version 22.05 of the library, but MATLAB don´t have the option of change to a update version of this. What can or should I do?

[Prasanth Sunkara]

Could you please share more info that will help us to reproduce this issue?

1) Which OS are you running on the Pi? Run "cat /etc/os-release" on pi's terminal

2) OS bitness. Run "getconf LONG_BIT"

3) Which version of MATLAB are you using?

[Yusuke Konno]

I also encounterd this problem. My environment is:

Raspberry pi 5,

raspberry os (bookworm),

latest hardware support package.

Answer 1

[Prasanth Sunkara]

ARM Compute Library is not necessary from R2024b onwards.

The Simulink/MATLAB generated C code for Deep learning models is already optimized to perform well on Raspberry Pi.

You do not have to select ARM Compute library in the Configuration parameters of Simulink model when working with DL models. Set the Deep learning Library parameter to "None" and proceed with deploying the model.