Requirements of BlumTelegramWindow

[dinhsang1999]

I used the code to my account on my device. But these click is wrong and alway click bomb. I think the code need one minimize device`s requirement.
My device do not have GPU. So I need your advise?

And How to your train your model, maybe I can train model on my device, it will be get more accurancy.

[behroozUAC]

I used the code to my account on my device. But these click is wrong and alway click bomb. I think the code need one minimize device`s requirement. My device do not have GPU. So I need your advise?

And How to your train your model, maybe I can train model on my device, it will be get more accurancy.

You definitely need a GPU to run the model smoothly, without delays or errors.
Otherwise, you’ll need a CPU with at least 14 cores and 20 threads to ensure it runs error free.

[Alirezaafshar20]

I have already tried installing the ultralytics package using pip, but it seems that the installation is unsuccessful. I also ensured that all dependencies are up to date, but the issue persists.

Could you please assist me in resolving this issue? Any guidance or steps to ensure that the module is installed correctly would be greatly appreciated.

[behroozUAC]

I have already tried installing the ultralytics package using pip, but it seems that the installation is unsuccessful. I also ensured that all dependencies are up to date, but the issue persists.

Could you please assist me in resolving this issue? Any guidance or steps to ensure that the module is installed correctly would be greatly appreciated.

Hi ,what issue ? Please specify the exact error

[dinhsang1999]

Thank your solution, it work. I ran the code with GPU.

[dinhsang1999]

Sir, could you give me some tutor to easy train the model which detect some project like this. How to label data or create data to train?

[behroozUAC]

Sir, could you give me some tutor to easy train the model which detect some project like this. How to label data or create data to train?

First, I made sure Python 3.7 or higher was installed, and then I grabbed the YOLO framework by running:

pip install ultralytics

Next, I gathered some Blum images I wanted to detect and labeled them using LabelImg. I made sure the labels were in YOLO format:

class_id x_center y_center width height

Here, the class_id is the object type, and the rest are box coordinates normalized to the image size.

After organizing everything into folders like this:

dataset/
├── images/
│   ├── train/ (training images)
│   ├── val/ (validation images)
├── labels/
│   ├── train/ (training labels)
│   ├── val/ (validation labels)

I wrote a quick script to start training:

from ultralytics import YOLO

# Load the model
model = YOLO('yolov8n.pt')  # Using the nano version for speed

# Train the model
model.train(
    data='data.yaml',  # Path to your data config
    epochs=50,
    imgsz=640,
    batch=16
)

The model logs its progress and saves the best version automatically. After training, I tested it on some new images:

model.predict(source='path_to_test_images/', save=True)

For deployment, exporting the model to ONNX or another format was a breeze:

model.export(format='onnx')