training on my own dataset

I have my own dataset with the landmark annotations like .pts file, txt file or csv file, how can I convert my annotation consistent with your dataset annotation in .t7? Look forward to your answer.

Look at the dataset loading and preprocessing code, likely in scripts or data/ directories, to understand the .t7 format. Check if there are existing conversion utilities or documentation. The goal is to write a script that transforms .pts, .txt, or .csv annotations into the .t7 format used by the training pipeline.

Written by the indexing model from the issue text.