Hi, thanks for your benchmark work and the open source code. I have a question about the Hybrid-training part of the code.
In the code, adversarial point clouds produced by pgd,drop and add are mixed together as an defense dataset. However, the num_points of pgd.npz, Drop.npz and Add.npz are different, which leads to code errors that the tensors of the dataset are not aligned. I'd appreciate it if you could answer my question. Thank you so much.
Hi, thanks for your benchmark work and the open source code. I have a question about the Hybrid-training part of the code.
In the code, adversarial point clouds produced by pgd,drop and add are mixed together as an defense dataset. However, the num_points of pgd.npz, Drop.npz and Add.npz are different, which leads to code errors that the tensors of the dataset are not aligned. I'd appreciate it if you could answer my question. Thank you so much.