Rotational augmentation

[SunshineDHW]

Here’s the English translation of your request:

"In the title of Table 1, you mentioned '† indicates that using simple rotational augmentation [11] for training.' How is this rotational augmentation implemented in the code? Additionally, I noticed that you also applied rotational augmentation to Point-MAE in Table 1. How is that part implemented? Thank you very much!"

[LMD0311]

Please refer to train_transforms in runner_finetune.py https://github.com/LMD0311/PointMamba/blob/da441cc374eed89d20be85887f08dd0bec3d619a/tools/runner_finetune.py#L22

[SunshineDHW]

Thank you very much!