lanyeeee
lanyeeee
same problem here in my case, in each epoch, firt load something into memory before the train, then memory is freed at the end of the train, load something into...
```python from ultralytics import YOLO if __name__ == '__main__': model = YOLO('weights/yolov8s.pt') model.train(data='data/mycoco.yaml', epochs=700, imgsz=640, project='runs') ``` This is the code I train the model
> So right now, for me, this is isolated to windows. > > Testing on Linux Mint, the time from stopping an epoch, and starting a new epoch is nearly...
> @BraunGe I think that may caused by your CPU. It takes time to load images for next epoch. What is your workers number? I encountered the same issue before,...
this problem not fixed until today :(
https://github.com/ultralytics/ultralytics/issues/509#issuecomment-1399660901 I just tested in **Linux Mint** and it works well. Maybe this issue only occurs in **Windows**
> @lanyeeee I haven't seen long load times in Google Colab or Docker on our server, but I don't have a windows machine to test that os. > > BTW...
I can't believe it's almost 2024 and this hasn't been fixed :(
having the same problem.
> You can just add a listener to the selection-area by yourself and clear the selection if there hasn't been a selection between the `pointerdown` and `pointerup` events, or something...