Hi, In the README, it mentions that the model you provided is simplified. Is the model trained by only segmentation mask?  thanks for helping!