assistance request: SynoAI utilizing CodeProjectAI

[JoeyJoe6]

I am attempting to utilize synoAI with codeprojectAI (GPU version) and having some small issues. Below is some background info on my setup, another set of eyes to comment would be greatly appreciated. This is a new install i am testing using a DVA1622. I currently am utilizing synoAI with deepstack (cpu version) on xpenology DSM 6.2 with great success.

SynoAI info:
Docker installed (latest version) on DVA1622 7.1 - camera Trigger appears to communicate fine with synoai

part of synoAI appsettings.json
"AI": {
"Type": "CodeProjectAIServer",
"Url": "http://192.168.42.249:32168"

SynoAI STdout (this output is being generated with a trigger "test" from the action rule I created)
2023-02-05T15:01:13.310832895Z,stdout, cam1: Removing running camera block.
2023-02-05T15:01:13.310784600Z,stdout, [40m [37mdbug [39m [22m [49m: SynoAI.Controllers.CameraController[0]
2023-02-05T15:01:13.310733194Z,stdout, cam1: Failed to get get predictions.
2023-02-05T15:01:13.310658677Z,stdout, [41m [30mfail [39m [22m [49m: SynoAI.Controllers.CameraController[0]
2023-02-05T15:01:13.310466468Z,stdout, cam1: DeepStackAI: Failed with unknown error.
2023-02-05T15:01:13.310409073Z,stdout, [40m [1m [33mwarn [39m [22m [49m: SynoAI.Services.AIService[0]

cam1: DeepStackAI: Responded with {"error":"The request timed out (#reqid {request.reqid})","success":false,"code":-1,"inferenceMs":0,"processMs":0,"analysisRoundTripMs":0}.

2023-02-05T15:01:13.310109318Z,stdout, [40m [37mdbug [39m [22m [49m: SynoAI.Services.AIService[0]
2023-02-05T15:01:13.268952784Z,stdout, cam1: DeepStackAI: POSTing image with minimum confidence of 0.45 (45%) to http://192.168.42.249:32168/v1/vision/detection.
2023-02-05T15:01:13.268789685Z,stdout, [40m [37mdbug [39m [22m [49m: SynoAI.Services.AIService[0]
2023-02-05T15:01:13.268691155Z,stdout, cam1: Snapshot 1 of 1 received at EVENT TIME 110ms.
2023-02-05T15:01:13.268610449Z,stdout, [40m [32minfo [39m [22m [49m: SynoAI.Controllers.CameraController[0]
2023-02-05T15:01:13.268462784Z,stdout, cam1: Snapshot received in 109ms.
2023-02-05T15:01:13.268344075Z,stdout, [40m [32minfo [39m [22m [49m: SynoAI.Controllers.CameraController[0]
2023-02-05T15:01:13.267785613Z,stdout, cam1: Reading snapshot

CodeProjectAI Info:
Windows 10 VM created in Proxmox
Full install, utilizing GPU and CUDA (server is operating with no errors)
Installed modules : Objection Detection (YOLOv5 6.2) GPU (CUDA) enabled

Started CodeProjectAI via CMDline with the below VISION-DETECTION=True flag
https://static.codeproject.com/AI/docs/why/module_configuration.html
sc stop "CodeProject.AI Server"
sc start "CodeProject.AI Server" --VISION-DETECTION=True
Listening port: 32168? (to send images from synoai)

Active Connections from Windows OS hosting CodeProjectAI (not a Docker, windows 10)
Proto Local Address Foreign Address State
TCP 0.0.0.0:135 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:445 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:3389 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:5000 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:5040 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:32168 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49664 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49665 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49666 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49667 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49668 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49669 CodeProjectAI:0 LISTENING
TCP 0.0.0.0:49725 CodeProjectAI:0 LISTENING
TCP 127.0.0.1:32168 CodeProjectAI:49825 TIME_WAIT
TCP 127.0.0.1:32168 CodeProjectAI:49835 TIME_WAIT
TCP 192.168.42.249:139 CodeProjectAI:0 LISTENING
TCP [::]:135 CodeProjectAI:0 LISTENING
TCP [::]:445 CodeProjectAI:0 LISTENING
TCP [::]:3389 CodeProjectAI:0 LISTENING
TCP [::]:5000 CodeProjectAI:0 LISTENING
TCP [::]:32168 CodeProjectAI:0 LISTENING
TCP [::]:49664 CodeProjectAI:0 LISTENING
TCP [::]:49665 CodeProjectAI:0 LISTENING
TCP [::]:49666 CodeProjectAI:0 LISTENING
TCP [::]:49667 CodeProjectAI:0 LISTENING
TCP [::]:49668 CodeProjectAI:0 LISTENING
TCP [::]:49669 CodeProjectAI:0 LISTENING
TCP [::]:49725 CodeProjectAI:0 LISTENING
TCP [::1]:32168 CodeProjectAI:49824 TIME_WAIT
TCP [::1]:32168 CodeProjectAI:49826 TIME_WAIT
TCP [::1]:32168 CodeProjectAI:49828 TIME_WAIT
TCP [::1]:32168 CodeProjectAI:49834 TIME_WAIT
TCP [::1]:32168 CodeProjectAI:49836 TIME_WAIT
TCP [::1]:32168 CodeProjectAI:59966 TIME_WAIT
UDP 0.0.0.0:123 :
UDP 0.0.0.0:500 :
UDP 0.0.0.0:3389 :
UDP 0.0.0.0:4500 :
UDP 0.0.0.0:5050 :
UDP 0.0.0.0:5353 :
UDP 0.0.0.0:5355 :
UDP 127.0.0.1:1900 :
UDP 127.0.0.1:56885 :
UDP 127.0.0.1:56946 :
UDP 192.168.42.249:137 :
UDP 192.168.42.249:138 :
UDP 192.168.42.249:1900 :
UDP 192.168.42.249:56884 :
UDP [::]:123 :
UDP [::]:500 :
UDP [::]:3389 :
UDP [::]:4500 :
UDP [::]:5353 :
UDP [::]:5355 :
UDP [::1]:1900 :
UDP [::1]:56883 :

Testing:
Able to ping from synoAI to CodeProjectAI server (tested via bash under docker/synoai/terminal)
Able to ping from CodeProjectAI to SynoAI docker container
Fire wall disabled on both DSM hosting synoAI docker and CodeProjectAI Windows OS
Attempted to use other ports within synoAI "Type": "CodeProjectAIServer", "Url": "http://192.168.42.249:83", 5000, 5040

[djdd87]

Can you provide your config, please?

[JoeyJoe6]

I just "##" out passwords, but kept it simple for testing.

{
"Logging": {
"LogLevel": {
"Default": "Information",
"Microsoft": "Warning",
"Microsoft.Hosting.Lifetime": "Information"
}
},

"Url": "http://192.168.42.248:5000",
"User": "synoai",
"Password": "##",

"MinSizeX": 100,
"MinSizeY": 100,

"DaysToKeepCaptures": 2,

"AI": {
"Type": "CodeProjectAIServer",
"Url": "http://192.168.42.249:32168"
},

"Notifiers": [
{
"Type": "Email",
"Sender": "[email protected]",
"Destination": "[email protected]",
"Host": "192.168.42.15",
"Port": 25,
"Username": "event",
"Password": "##",
"Encryption": "None"
}
],

"Cameras": [
{
"Name": "cam1",
"Types": [ "Person, bicycle, car, motorbike, bus, truck, boat, cat, dog, horse" ],
"Threshold": 45,
"MinSizeX": 150,
"MinSizeY": 150
},

]

}

SynoAI log when test trigger sent:
cam1: The camera is currently running; no other camera requests will be processed while this request is ongoing.
info: SynoAI.Controllers.CameraController[0]
Captures Clean Up: Cleaning up images older than 2 day(s).
info: SynoAI.Controllers.CameraController[0]
cam1: Snapshot 1 of 1 requested at EVENT TIME 0ms.
dbug: SynoAI.Services.SynologyService[0]
cam1: Found with Synology ID '1'.
dbug: SynoAI.Services.SynologyService[0]
cam1: Taking snapshot from 'webapi/entry.cgi?version=9&id=1&api="SYNO.SurveillanceStation.Camera"&method=GetSnapshot&profileType=1'.
info: SynoAI.Services.SynologyService[0]
cam1: Taking snapshot
dbug: SynoAI.Services.SynologyService[0]
cam1: Reading snapshot
info: SynoAI.Controllers.CameraController[0]
cam1: Snapshot received in 92ms.
info: SynoAI.Controllers.CameraController[0]
cam1: Snapshot 1 of 1 received at EVENT TIME 93ms.
dbug: SynoAI.Services.AIService[0]
cam1: DeepStackAI: POSTing image with minimum confidence of 0.45 (45%) to http://192.168.42.249:32168/v1/vision/detection.
dbug: SynoAI.Services.AIService[0]
cam1: DeepStackAI: Responded with {"success":false,"error":"Error occured on the server","code":500,"analysisRoundTripMs":109}.
warn: SynoAI.Services.AIService[0]
cam1: DeepStackAI: Failed with unknown error.
fail: SynoAI.Controllers.CameraController[0]
cam1: Failed to get get predictions.
dbug: SynoAI.Controllers.CameraController[0]
cam1: Removing running camera block.

codeproject syslog:
09:43:27:Client request 'detect' in the queue (...5d5443)
09:43:27:Request 'detect' dequeued for processing (...5d5443)
09:43:27:Object Detection (YOLOv5 6.2): Retrieved objectdetection_queue command
09:43:27:Response received (...5d5443)
09:43:27:Object Detection (YOLOv5 6.2): [Exception] : Traceback (most recent call last):
09:43:27:Object Detection (YOLOv5 6.2): Queue and Processing Object Detection (YOLOv5 6.2) command 'detect' (...5d5443) took 16ms

**Seems to be an instant failure on SynoAI's end. Is there a timeout value I can modify to wait for a response longer?

[djdd87]

Try and define your types properly:

"Types": [ "Person", "bicycle", "car", "motorbike", "bus", "truck", "boat", "cat", "dog", "horse" ],

[JoeyJoe6]

Thanks for the eyes. Modified the types, but this did not help.
Final Solution: SynoAI did not like working with SS CMS. I noticed that I forget to mention I was using CMS to control my DMS's. I installed SynoAI on each of the SS setups and modified appsettings for each instance. ALL IS WORKING NOW :-).. THANK YOU for a great product!!

[djdd87]

I've no idea what that even is! So well done for sorting it.

[JoeyJoe6]

I also see this popup on DMS7 interface.