Hi everyone, I want to know if it is possible to train only one task at a time on a custom dataset. For example, I only want to train the lane segmentation part, but I don't have an object detection dataset. However, from the code, it looks like it also checks for the presence of an object detection dataset. How do I modify the code to bypass this check?
Hi everyone, I want to know if it is possible to train only one task at a time on a custom dataset. For example, I only want to train the lane segmentation part, but I don't have an object detection dataset. However, from the code, it looks like it also checks for the presence of an object detection dataset. How do I modify the code to bypass this check?