I’m training a custom dataset for multiple-class object detection and segmentation, consisting of over 1500 images. However, I’m encountering an issue where the training process is taking an excessively long time without showing any error messages. The training runs successfully up to 100 epochs but then repeats the same operation without making progress.