We attach the code for training and model loading in our submission. Please note that these codes may not be executable as the dataset has not been publicly released within the review process. However, we will publicly release the dataset, code, and trained model. Additionally, we included three random samples for our dataset.