CodeProject.AI Version 2.0

MikeLud1 · Aug 5, 2023

Mad_Marik said:
I don't see the location exactly that you show, this is what I see

It looks like it never download for some reason. Maybe give it a try again.

tofu · Aug 5, 2023

Daniel15 said:
Frigate uses "tiny" models and a lot of people still find value in it. Some even use Frigate and BI concurrently for whatever reason.

ignorance is bliss, i guess?

Mad_Marik · Aug 5, 2023

MikeLud1 said:
It looks like it never download for some reason. Maybe give it a try again.

No joy, same error, restarted service, and restarted computer. Should try full uninstall and reinstall?

When I tried the restart of the service, it did hang a bit, then failed to restart, so I had to manually start it (which it did fine) but the error was still there.

MikeLud1 · Aug 5, 2023

Mad_Marik said:
No joy, same error, restarted service, and restarted computer. Should try full uninstall and reinstall?

When I tried the restart of the service, it did hang a bit, then failed to restart, so I had to manually start it (which it did fine) but the error was still there.

Is the any more errors in the log then Error in Install TrainingYoloV5: Call failed.

Mad_Marik · Aug 5, 2023

MikeLud1 said:
Is the any more errors in the log then Error in Install TrainingYoloV5: Call failed.

Negative, just that, everything else is running fine. Just tried to install that module and it popped that error.

MikeLud1 · Aug 5, 2023

Mad_Marik said:
Negative, just that, everything else is running fine. Just tried to install that module and it popped that error.

You may need to refresh the webpage by doing a Ctrl+F5. After it refreshes try installing the module.

Mad_Marik · Aug 5, 2023

MikeLud1 said:
You may need to refresh the webpage by doing a Ctrl+F5. After it refreshes try installing the module.

Negative, same error sir. Again, its not a big deal, nothing for you to lose sleep over, just giving you info.

wpiman · Aug 6, 2023

Mike...

I pulled the latest docker image on my Orange Pi 5. I can see the RKNN object detection module. However, when I go to start it, I get an error.

Are you running it natively? Have you tried it in docker? I am wondering if I need to pass through any device....

10:06:48:Error trying to start Object Detection (YOLOv5 RKNN) (objectdetection_fd_rknn_adapter.py)
10:06:48:An error occurred trying to start process '/app/modules/ObjectDetectionYoloRKNN/bin/linux/python39/venv/bin/python3' with working directory '/app/modules/ObjectDetectionYoloRKNN'. No such file or directory
10:06:48: at System.Diagnostics.Process.ForkAndExecProcess(ProcessStartInfo startInfo, String resolvedFilename, String[] argv, String[] envp, String cwd, Boolean setCredentials, UInt32 userId, UInt32 groupId, UInt32[] groups, Int32& stdinFd, Int32& stdoutFd, Int32& stderrFd, Boolean usesTerminal, Boolean throwOnNoExec)
at System.Diagnostics.Process.StartCore(ProcessStartInfo startInfo)
at CodeProject.AI.API.Server.Frontend.ModuleProcessServices.StartProcess(ModuleConfig module)
10:06:48

lease check the CodeProject.AI installation completed successfully
10:06:48:

wpiman · Aug 6, 2023

root@ec1b83340bc4:/app/modules/ObjectDetectionYoloRKNN# cat install.log
2023-08-06 14:01:00: Hi Docker! We will disable shared python installs for downloaded modules
2023-08-06 14:01:00: No schemas installed
2023-08-06 14:01:00: (No schemas means: we can't detect if you're in light or dark mode)
2023-08-06 14:01:00: Setting up CodeProject.AI Development Environment
2023-08-06 14:01:00: ======================================================================
2023-08-06 14:01:00: CodeProject.AI Installer
2023-08-06 14:01:00: ======================================================================
2023-08-06 14:01:00: Checking GPU support
2023-08-06 14:01:00: CUDA Present...No
2023-08-06 14:01:00: Allowing GPU Support: Yes
2023-08-06 14:01:00: Allowing CUDA Support: Yes
2023-08-06 14:01:00: General CodeProject.AI setup
2023-08-06 14:01:00: Creating Directories...Done
2023-08-06 14:01:00: Installing module ObjectDetectionYoloRKNN
2023-08-06 14:01:00: /app/modules/ObjectDetectionYoloRKNN/install.sh: line 22: /sys/firmware/devicetree/base/model: No such file or directory
2023-08-06 14:01:00: ObjectDetection (Fast Deploy RKNN) can only be installed on Orange Pi devices
Installer exited with code 0
root@ec1b83340bc4:/app/modules/ObjectDetectionYoloRKNN# ls /sys/firmware/
root@ec1b83340bc4:/app/modules/ObjectDetectionYoloRKNN#

MikeLud1 · Aug 6, 2023

wpiman said:
Mike...

I pulled the latest docker image on my Orange Pi 5. I can see the RKNN object detection module. However, when I go to start it, I get an error.

Are you running it natively? Have you tried it in docker? I am wondering if I need to pass through any device....

View attachment 169447

10:06:48:Error trying to start Object Detection (YOLOv5 RKNN) (objectdetection_fd_rknn_adapter.py)
10:06:48:An error occurred trying to start process '/app/modules/ObjectDetectionYoloRKNN/bin/linux/python39/venv/bin/python3' with working directory '/app/modules/ObjectDetectionYoloRKNN'. No such file or directory
10:06:48: at System.Diagnostics.Process.ForkAndExecProcess(ProcessStartInfo startInfo, String resolvedFilename, String[] argv, String[] envp, String cwd, Boolean setCredentials, UInt32 userId, UInt32 groupId, UInt32[] groups, Int32& stdinFd, Int32& stdoutFd, Int32& stderrFd, Boolean usesTerminal, Boolean throwOnNoExec)
at System.Diagnostics.Process.StartCore(ProcessStartInfo startInfo)
at CodeProject.AI.API.Server.Frontend.ModuleProcessServices.StartProcess(ModuleConfig module)
10:06:48lease check the CodeProject.AI installation completed successfully
10:06:48:

Below is how I run the docker container. One note you need to create "/etc/codeproject/ai" and "/app/modules" folders under your user folder first before running

Code:

docker run --restart=always --name CodeProject.AI-Server-arm64-2.1.10 -d -p 32168:32168 \
 -t -i --privileged -v /dev/bus/usb:/dev/bus/usb \
 --mount type=bind,source=./etc/codeproject/ai,target=/etc/codeproject/ai \
 --mount type=bind,source=./opt/codeproject/ai,target=/app/modules \
   codeproject/ai-server:arm64-2.1.10

MikeLud1 · Aug 6, 2023

wpiman said:
root@ec1b83340bc4:/app/modules/ObjectDetectionYoloRKNN# cat install.log
2023-08-06 14:01:00: Hi Docker! We will disable shared python installs for downloaded modules
2023-08-06 14:01:00: No schemas installed
2023-08-06 14:01:00: (No schemas means: we can't detect if you're in light or dark mode)
2023-08-06 14:01:00: Setting up CodeProject.AI Development Environment
2023-08-06 14:01:00: ======================================================================
2023-08-06 14:01:00: CodeProject.AI Installer
2023-08-06 14:01:00: ======================================================================
2023-08-06 14:01:00: Checking GPU support
2023-08-06 14:01:00: CUDA Present...No
2023-08-06 14:01:00: Allowing GPU Support: Yes
2023-08-06 14:01:00: Allowing CUDA Support: Yes
2023-08-06 14:01:00: General CodeProject.AI setup
2023-08-06 14:01:00: Creating Directories...Done
2023-08-06 14:01:00: Installing module ObjectDetectionYoloRKNN
2023-08-06 14:01:00: /app/modules/ObjectDetectionYoloRKNN/install.sh: line 22: /sys/firmware/devicetree/base/model: No such file or directory
2023-08-06 14:01:00: ObjectDetection (Fast Deploy RKNN) can only be installed on Orange Pi devices
Installer exited with code 0
root@ec1b83340bc4:/app/modules/ObjectDetectionYoloRKNN# ls /sys/firmware/
root@ec1b83340bc4:/app/modules/ObjectDetectionYoloRKNN#

Also what image did you use for the OS, I am using Armbian 23.5 Jammy Gnome

Orange Pi 5 and 5B - Armbian

PCI support is broken on vendor kernel (6.1.xx) on versions higher then 24.8.1. Use this version and freeze kernel (armbian-config -> system) prior to updating OS. PCI works on other Rockchip 3588 devices, while fixing this on all represent expensive technical challenge which we will be able to...

www.armbian.com

wpiman · Aug 6, 2023

MikeLud1 said:
Also what image did you use for the OS, I am using Armbian 23.5 Jammy Gnome

Orange Pi 5 and 5B - Armbian

PCI support is broken on vendor kernel (6.1.xx) on versions higher then 24.8.1. Use this version and freeze kernel (armbian-config -> system) prior to updating OS. PCI works on other Rockchip 3588 devices, while fixing this on all represent expensive technical challenge which we will be able to...

www.armbian.com

Ubuntu...

mike@orangepi5:/dev$ cat /etc/os-release
PRETTY_NAME="Orange Pi 1.1.4 Jammy"
NAME="Ubuntu"
VERSION_ID="22.04"
VERSION="22.04.2 LTS (Jammy Jellyfish)"
VERSION_CODENAME=jammy
ID=ubuntu
ID_LIKE=debian
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
UBUNTU_CODENAME=jammy

Here is my docker-compose file.... I am not passing the USB on through. Is that for RKNN or for something else (CoralTPu?). I'll give that a shot...

I have those directories mounted...

services:
codeprojetai:
container_name: ai-server
build: .
image: codeproject/ai-server:arm64-2.1.10
volumes:
- ./data:/etc/codeproject/ai
- ./modules:/app/modules
ports:
- 32168:32168
restart: unless-stopped

MikeLud1 · Aug 6, 2023

wpiman said:
Is that for RKNN or for something else (CoralTPu?).

It is for RKNN

wpiman · Aug 6, 2023

Ok...

I passed that through. I see it in the container...

root@dffc6542bc73:/dev/bus/usb# ls
001 002 003 004 005 006 007 008
root@dffc6542bc73:/dev/bus/usb#

Latest YAML...

mike@orangepi5:~/Docker/aiServer$ cat docker-compose.yml

services:
codeprojetai:
container_name: ai-server
build: .
image: codeproject/ai-server:arm64-2.1.10
volumes:
- ./data:/etc/codeproject/ai
- ./modules:/app/modules
ports:
- 32168:32168
devices:
- /dev/bus/usb:/dev/bus/usb
privileged: true
restart: unless-stopped

Still having some issues. Going to blow away my directories and try from scratch....

MikeLud1 · Aug 6, 2023

wpiman said:
Ubuntu...

mike@orangepi5:/dev$ cat /etc/os-release
PRETTY_NAME="Orange Pi 1.1.4 Jammy"
NAME="Ubuntu"
VERSION_ID="22.04"
VERSION="22.04.2 LTS (Jammy Jellyfish)"
VERSION_CODENAME=jammy
ID=ubuntu
ID_LIKE=debian
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
UBUNTU_CODENAME=jammy

Here is my docker-compose file.... I am not passing the USB on through. Is that for RKNN or for something else (CoralTPu?). I'll give that a shot...

I have those directories mounted...

services:
codeprojetai:
container_name: ai-server
build: .
image: codeproject/ai-server:arm64-2.1.10
volumes:
- ./data:/etc/codeproject/ai
- ./modules:/app/modules
ports:
- 32168:32168
restart: unless-stopped

Also make sure you are installing V1.1 not V1.0 there was an install issue with V1.0.

wpiman · Aug 6, 2023

Interesting.. I just created clean directories and I see no version numbers at all there....

Mad_Marik · Aug 6, 2023

MikeLud1 said:
You may need to refresh the webpage by doing a Ctrl+F5. After it refreshes try installing the module.

This is all the server logs shows. Call failed

wpiman · Aug 6, 2023

Ok, so I went back to 2.1.9-- installed version 1.1 of the RKNN successfully... Then I moved to 2.1.10 and I ran the object detection.

Very fast.... Very nice! Not sure why I don't see version numbers in 2.1.10??!!

direkt · Aug 6, 2023

Is there a way to load custom models with the RPI docker image while using the Coral or TF-Lite option? I only have one in the drop down.

MikeLud1 · Aug 6, 2023

wpiman said:
Ok, so I went back to 2.1.9-- installed version 1.1 of the RKNN successfully... Then I moved to 2.1.10 and I ran the object detection.

Very fast.... Very nice! Not sure why I don't see version numbers in 2.1.10??!!

Try refreshing the page by doing a Ctrl+F5, I think it might fix it.

Search

Search

CodeProject.AI Version 2.0

MikeLud1

tofu

Getting the hang of it

Mad_Marik

n3wb

MikeLud1

Mad_Marik

n3wb

MikeLud1

Mad_Marik

n3wb

wpiman

Getting comfortable

wpiman

Getting comfortable

MikeLud1

MikeLud1

Orange Pi 5 and 5B - Armbian

wpiman

Getting comfortable

Orange Pi 5 and 5B - Armbian

MikeLud1

wpiman

Getting comfortable

MikeLud1

wpiman

Getting comfortable

Attachments

Mad_Marik

n3wb

Attachments

wpiman

Getting comfortable

direkt

n3wb

MikeLud1