Thank you for your excellent work on Point-SAM. I’m currently trying to run inference using your model on custom .ply point cloud data. While the training pipeline is well-documented, I noticed that the inference code is incomplete or not clearly provided in the repository.

Would it be possible to share a complete and minimal example of the inference pipeline (including data loading, preprocessing, prompt handling, and mask output) in the official GitHub repo?

This would be extremely helpful for the community trying to apply Point-SAM to real-world datasets.