I'm having a problem with X-anylabeling trying to use an algorithmic model to batch label images.

[JIAMENGL]

My x-anylabelling is run by the source code. The computer configuration is Windows, no GPU. I had problems with using the AI model, and as I saw in the first image, I didn't know how to modify it.

[CVHub520]

Hi @JIAMENGL,

Thanks for reporting this. It looks like this might be an issue that has been resolved in a more recent version of the code.

Could you please update your local repository to the latest commit? You can do this by running the following command in your terminal from the project directory:

git pull

After updating, please try running the application again.

If the error persists, please copy and paste the full log output from your terminal and upload a minimal reproduction video here. The complete traceback is essential for us to accurately diagnose the issue.

Thanks!

[JIAMENGL]

I saw the system indicating it was the latest version. I have attached an image that shows the error message from the terminal.

[CVHub520]

Thanks for providing the full log.

The key error message here is from OpenCV:

error: (-5: Bad argument) in function 'imdecode'
> Overload resolution failed:
>  - buf is not a numpy array, neither a scalar

This indicates that the program is failing to read one of your image files. This is typically caused by one of the following:

1. The file path or filename contains non-English (non-ASCII) characters.
2. The image file itself is corrupted or is not in a supported format.

Please check the folder containing your images and ensure that both the folder path and the individual filenames use only English letters and numbers.

As a quick test, you could try pointing the application to a new folder with a simple path (e.g., C:\test_images\) that contains a single, known-good image file. If that works, it will confirm the issue is with your original image paths or files.

[JIAMENGL]

I have confirmed that the naming of the image is correct, but I still cannot read the image data correctly ....

[CVHub520]

Thanks for your continued efforts. It seems the issue is persisting. Let's try a couple more things to isolate the cause.

The error might be related to the specific model you are using (damo_yolo) or how the images are being loaded.

1. Try a Different Model:
Could you please try switching to a different model from the dropdown list, for example, yolov8s? Then try running the prediction again.

2. Load Images Directly from the Command Line:
Instead of opening the folder from the GUI, please try launching the application and loading your images with a single command.

Open your terminal, navigate to the X-AnyLabeling directory, and run:

python anylabeling/app.py C:\path\to\your\images

Important:

• Replace C:\path\to\your\images with the full path to your image folder.
• Make sure the folder you specify contains only image files and no subfolders.

Please let us know if either of these steps resolves the error or changes the outcome.

[JIAMENGL]

Directly loading images from the command line still cannot solve the problem of reading images. In addition, I tried to obtain the YOLOV8S model, but it seems that I cannot obtain it due to network issues (please refer to the picture for details). My current network check has confirmed that there is no problem.

[JIAMENGL]

Thanks for your answer. I will try reinstalling x-anylabelling tomorrow or try other image annotation tools. Could it be that there was a problem configuring the virtual environment? I followed the user documentation step by step ......emmmm

[CVHub520]

We haven’t encountered this issue on our side yet — it’s quite unusual. You might try running the pre-built executable version to help rule out any problems related to your local environment.

Also, there are many excellent annotation tools available in the community. Feel free to explore and see if any of them meet your needs!