How to build CodeProject.AI on NVIDIA Jetson Orin Nano?

[styxol]

I tried to build CodeProject.AI on Jetson Orin Nano following steps from Installing CodeProject.AI Server on a Jetson Nano as a Development environment. It fails on step 8.
It looks like all dependencies are installed, but the script failed to validate. Why might the validation be failing?

For example, if I try to install Pillow manually using pip install Pillow, it tells me:

Requirement already satisfied: Pillow in /usr/lib/python3/dist-packages (9.0.1)

`======================================================================

               CodeProject.AI Installer                           

======================================================================

861.00 GiB of 936.03 GiB available on Jetson (linux ubuntu arm64 - jetson)
Installing xz-utils...

General CodeProject.AI setup

Setting permissions on runtimes folder...done
Setting permissions on general downloads folder...done
Setting permissions on module asset downloads folder...done
Setting permissions on modules folder...done
Setting permissions on models folder...done
Setting permissions on persisted data folder...done

GPU support

CUDA (NVIDIA) Present: Yes (CUDA 12.6.68, cuDNN 9.3.0)
ROCm (AMD) Present: No
MPS (Apple) Present: No

Processing CodeProject.AI SDK

Adding toolchain repo...Already added
Searching for apt-utils...All good.
Searching for ca-certificates gnupg libc6-dev libfontc...All good.
Searching for ffmpeg libsm6 libxext6 mesa-utils curl r...All good.
Checking for .NET 9.0.0 aspnetcore...All good. .NET aspnetcore is 9.0.1
.NET aspnetcore 9.0.1 is present.
Checking for .NET 9.0.100 sdk...All good. .NET sdk is 9.0.102
.NET sdk 9.0.102 is present.

Processing CodeProject.AI Server

Nothing to do here.

Processing Included CodeProject.AI Server Modules

Reading module settings.......done
Processing module ObjectDetectionYOLOv5-6.2 1.10.0 (Internal)

This module cannot be installed on this system

Reading module settings.......done
Processing module ObjectDetectionYOLOv5Net 1.14.0 (Internal)

Building project...
Downloading Nuget packages... already exists...Expanding...done.
Moving contents of ObjectDetectionYOLOv5NetNugets.zip to LocalNugets...done.
Downloading YOLO ONNX models... already exists...Expanding...done.
Moving contents of yolonet-models.zip to assets...done.
Downloading Custom YOLO ONNX models... already exists...Expanding...done.
Moving contents of yolonet-custom-models.zip to custom-models...done.
Scanning modulesettings for downloadable models...No models specified
Self test: Self-test passed
Module setup time 00:00:21

Processing External CodeProject.AI Server Modules

No external modules found

Module setup Complete

Processing demo clients

Python 3.8 is already installed
Ensuring PIP in base python install...done
Upgrading PIP in base python install...done
Virtual Environment already present
Checking for Python 3.8...(Found Python 3.8.20) All good
Upgrading PIP in virtual environment...done
Installing updated setuptools in venv...done
Searching for python3-pip...All good.
Ensuring PIP compatibility...done
Python packages will be specified by requirements.txt

• Installing Pillow, the Python image library...(❌ failed check) done
• Installing imutils, the image utilities library...(❌ failed check) done
• Installing OpenCV, the Computer Vision library for Python...(❌ failed check) done
• Installing NumPy, a package for scientific computing...(❌ failed check) done

Processing demo modules

Reading module settings.......done
Processing module DotNetLongProcess 1.1.0 (Demo)

Building project...
Scanning modulesettings for downloadable models...No models specified
Self test: Self-test passed
Module setup time 00:00:09

Reading module settings.......done
Processing module DotNetSimple 1.1.0 (Demo)

Building project...
Downloading YOLO ONNX models... already exists...Expanding...done.
Moving contents of objectdetection-coco-yolov8-onnx-m.zip to assets...done.
Scanning modulesettings for downloadable models...No models specified
Self test: Self-test passed
Module setup time 00:00:12

Reading module settings.......done
Processing module PythonLongProcess 1.0.0 (Demo)

Installing Python 3.9
Python 3.9 is already installed
Ensuring PIP in base python install...done
Upgrading PIP in base python install...done
Virtual Environment already present
Checking for Python 3.9...(Found Python 3.9.20) All good
Upgrading PIP in virtual environment...done
Installing updated setuptools in venv...done
Installing Python packages for Python Long Process Demo
Installing GPU-enabled libraries: If available
Searching for python3-pip...All good.
Ensuring PIP compatibility...done
Python packages will be specified by requirements.txt

• Installing the CodeProject.AI SDK...(❌ failed check) done
  Scanning modulesettings for downloadable models...No models specified
  Self test: Traceback (most recent call last):
  File "/home/user/codeprojectai/CodeProject.AI-Server/src/demos/modules/PythonLongProcess/long_process_demo_adapter.py", line 9, in
  from codeproject_ai_sdk import RequestData, ModuleRunner, JSON
  ModuleNotFoundError: No module named 'codeproject_ai_sdk'
  Self-test failed
  Module setup time 00:00:47

Reading module settings.......done
Processing module PythonSimple 1.0.0 (Demo)

Installing Python 3.8
Python 3.8 is already installed
Ensuring PIP in base python install...done
Upgrading PIP in base python install...done
Virtual Environment already present
Checking for Python 3.8...(Found Python 3.8.20) All good
Upgrading PIP in virtual environment...done
Installing updated setuptools in venv...done
Downloading Standard YOLO models... already exists...Expanding...done.
Moving contents of models-yolo8-pt.zip to assets...done.
Installing Python packages for Python Simple Module demo
Installing GPU-enabled libraries: If available
Searching for python3-pip...All good.
Ensuring PIP compatibility...done
Python packages will be specified by requirements.linux.cuda.txt

• Installing NumPy, a package for scientific computing...(❌ failed check) done

• Installing PyTorch, an open source machine learning framework...(❌ failed check) done

• Installing TorchVision, for working with computer vision models...(❌ failed check) done

• Installing Ultralytics package for object detection in images...(❌ failed check) done

• Installing the CodeProject.AI SDK...(❌ failed check) done
  Scanning modulesettings for downloadable models...No models specified
  Executing post-install script for Python Simple Module demo
  Applying Ultralytics patch
  cp: cannot create regular file '/home/user/codeprojectai/CodeProject.AI-Server/src/demos/modules//PythonSimple/bin/ubuntu/python38/venv/lib/python3.8/site-packages//ultralytics/nn/.': No such file or directory
  Module setup time 00:01:47

          Setup complete                                        
  

Total setup time 00:04:50

SETUP FAILED:

• [Demo: PythonLongProcess @ /home/user/codeprojectai/CodeProject.AI-Server/src/demos/modules//PythonLongProcess] Self-test failed

• [Demo: PythonSimple @ /home/user/codeprojectai/CodeProject.AI-Server/src/demos/modules//PythonSimple] Error running post-install script`