|
Hey Chris,
Many thanks
Unfortunately i am still getting the exact same behaviour (2.3.3)
14:44:35:Started Object Detection (Coral) module
14:44:41:objectdetection_coral_adapter.py: Unable to load OpenCV or numpy modules. Only using PIL.
14:44:41:objectdetection_coral_adapter.py: Supporting multiple Edge TPUs
Then it falls back to using CPU (TF-Lite).
Same issue when trying to run pip command once I've activated the python environment also.
Is there a way I can install 2.3.0 to confirm everything is back okay on that version?
|
|
|
|
|
We released Coral 2.3.4 yesterday. Does that change anything? After uninstall module, then reinstall?
Thanks,
Sean Ewington
CodeProject
|
|
|
|
|
Still the same thing:
So I disabled Blue Iris and started it up, got the "Unable to load OpenCV or numpy modules. Only using PIL." message straight away, but it still showed it was using the TPU.
Left this for 8 mins and it continued to show TPU. The minute I started Blue Iris and it sent something to be analysed (14:45) it flips it to CPU.
15:37:37:Started Object Detection (Coral) module
15:37:43:objectdetection_coral_adapter.py: Unable to load OpenCV or numpy modules. Only using PIL.
15:37:43:objectdetection_coral_adapter.py: Supporting multiple Edge TPUs
15:45:20:Response rec'd from Object Detection (Coral) command 'detect' (...8f6fd0) [''] took 57ms
15:45:20:Response rec'd from Object Detection (Coral) command 'detect' (...6a3edb) ['No objects found'] took 90ms
|
|
|
|
|
"Unable to load OpenCV or numpy modules. Only using PIL" is just info. Ignore.
cheers
Chris Maunder
|
|
|
|
|
Is there anyway I can download 2.3.0? I believe this was the last one that worked correctly, just to see everything is okay my end?
|
|
|
|
|
There was no 2.3.0, but here's 2.3.1[^]
- Remove the \bin folder in the Coral module's folder
- Unzip this zip into the Coral module's folder
- open a terminal in that folder, and hit ....\setup
Be patient, and it should install.
cheers
Chris Maunder
|
|
|
|
|
Hi Chris,
Many thanks for the reply, is the version before 2.3.1 available? 2.3.1 seemed to be when things started to go wrong.
Thanks a lot
|
|
|
|
|
Try 2.2.2
cheers
Chris Maunder
|
|
|
|
|
Hi Chris,
So far so good on 2.2.2. Had to downgrade numpy from 2.0 to 1.21.6, but after that it started straight up and has stayed using TPU since.
|
|
|
|
|
yeah 2.2.2 was the last stable version for me
|
|
|
|
|
Yep 2.2.2 has been rock solid for me since last night now
|
|
|
|
|
Yep 2.2.2 rock solid again. Let me know if you need anything to help
|
|
|
|
|
At least one other person is reporting that 2.2.2[^] is working fine for them.
Thanks,
Sean Ewington
CodeProject
|
|
|
|
|
Just performed the downgrade to 2.2.2 under Windows. When starting the module the following failure messages are popping up. What can I do?
Quote: 16:37:26:Started Object Detection (Coral) module
16:37:26:objectdetection_coral_adapter.py: Traceback (most recent call last):
16:37:26:objectdetection_coral_adapter.py: File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionCoral\objectdetection_coral_adapter.py", line 10, in
16:37:26:objectdetection_coral_adapter.py: from module_runner import ModuleRunner
16:37:26:objectdetection_coral_adapter.py: File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionCoral\../../SDK/Python\module_runner.py", line 30, in
16:37:26:objectdetection_coral_adapter.py: import aiohttp
16:37:26:objectdetection_coral_adapter.py: ModuleNotFoundError: No module named 'aiohttp'
|
|
|
|
|
Unfortunately, when I downgrade the coral module, I get the message "This module cannot be installed on this system". Am I missing something?
1. I removed the bin folder -> C:\Program Files\CodeProject\AI\modules\ObjectDetectionCoral\bin
2. I copy/overwrite the content from 2.3.1 (i also try the other version below) in the \Program
Files\CodeProject\AI\modules\ObjectDetectionCoral folder.
3. then io run the setup.bat script from C:\Program Files\CodeProject\AI\setup.bat
has anyone seen where i made the mistake?
|
|
|
|
|
Have you loaded the batch file from within the module folder? The command should be
..\..\setup.bat
|
|
|
|
|
...that was my the problem... thx a lot but i run into the same issue as the comment (Jens_G) above.
I think I need to look further into where the problem might lie, but thanks for the help
|
|
|
|
|
The problem is after 60seconds of inactivity the script is unable to find the TPU module and is falling back on CPU detections and need to restart the module so it can continue to use TPU module
16:17:51:Response rec'd from Object Detection (Coral) command 'custom' (...8777ec) ['Found car, person, car...'] took 15ms
16:18:56:Client request 'custom' in queue 'objectdetection_queue' (...fcca9c)
16:18:56:Request 'custom' dequeued from 'objectdetection_queue' (...fcca9c)
16:18:56:Object Detection (Coral): Retrieved objectdetection_queue command 'custom'
16:18:56:objectdetection_coral_adapter.py: WARNING:root:No multi-TPU interpreters: Falling back to single-TPU/CPU detection
16:18:56:Response rec'd from Object Detection (Coral) command 'custom' (...fcca9c) ['Found car, person, car...'] took 44ms
16:19:12:objectdetection_coral_adapter.py: WARNING:root:No multi-TPU interpreters: Falling back to single-TPU/CPU listing
16:19:12:objectdetection_coral_adapter.py: WARNING:root:No multi-TPU interpreters: Falling back to single-TPU/CPU listing
16:20:12:objectdetection_coral_adapter.py: WARNING:root:No multi-TPU interpreters: Falling back to single-TPU/CPU listing
16:20:12:objectdetection_coral_adapter.py: WARNING:root:No multi-TPU interpreters: Falling back to single-TPU/CPU listing
|
|
|
|
|
Yup I saw that, everything worked fine on earlier version of module
|
|
|
|
|
Hi all,
I thought CodeProject included custom models with YOLOv5 6.2 in the past? I just upgraded my installation, and my v6.2 folder does not have any models or custom models (both actionnet and ipcam are missing).
However, my YOLOv5 v3.1 folder does seem to have models.
Should I move the models from v3.1 to v6.2, or do I need to manually download them?
I deleted the whole folder and reinstalled and am still having the same error. In the logs, I see the following:
22:45:29:Object Detection (YOLOv5 6.2): [FileNotFoundError] : Error during main_loop: Traceback (most recent call last):
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\codeproject_ai_sdk\module_runner.py", line 576, in main_loop
output = await callbacktask
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\lib\concurrent\futures\thread.py", line 57, in run
result = self.fn(*self.args, **self.kwargs)
File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionYOLOv5-6.2\detect_adapter.py", line 115, in process
response = self._list_models(self.opts.custom_models_dir)
File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionYOLOv5-6.2\detect_adapter.py", line 219, in _list_models
self.model_names = [entry.name[:-3] for entry in os.scandir(models_path)
FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\\Program Files\\CodeProject\\AI\\modules\\ObjectDetectionYOLOv5-6.2\\custom-models'
Thanks.
modified 25-Jun-24 22:48pm.
|
|
|
|
|
Thanks so much for the report. The custom models should still be in the C:\Program Files\CodeProject\AI\modules\ObjectDetectionYOLOv5-6.2\custom-models folder.
I suspect something else is going on here, but would you mind indulging me and try uninstalling, and reinsalling the module?
Also, if that doesn't work, would you please share your System Info tab from your CodeProject.AI Server dashboard? Then I'll sick the devs on this.
Thanks,
Sean Ewington
CodeProject
|
|
|
|
|
Hello,
I installed and reinstalled 2.6.2 and 2.6.5, both are missing the custom-models folder.
I also did a clean install and deleted the folder (Program Files\CodeProject) completely - same results.
|
|
|
|
|
We found the issue, updated the module to 1.9.2, and you should be able to download it now to resolve the problem.
Thanks,
Sean Ewington
CodeProject
|
|
|
|
|
Not sure how to roll back either... Everything was working with the last version of V8. It's only saying CPU now on the status screen. GPU is not being detected even though the info shows cuda device available
13:55:11:Started Object Detection (YOLOv8) module
13:55:11:detect_adapter.py: Traceback (most recent call last):
13:55:11:detect_adapter.py: File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionYOLOv8\detect_adapter.py", line 10, in
13:55:11:detect_adapter.py: from codeproject_ai_sdk import RequestData, ModuleRunner, LogMethod, JSON
13:55:11:detect_adapter.py: ModuleNotFoundError: No module named 'codeproject_ai_sdk'
13:55:11:Module ObjectDetectionYOLOv8 has shutdown
|
|
|
|
|
Can you please try reinstalling the module?
cheers
Chris Maunder
|
|
|
|