Why the new_scaling repeated ?

[ljh12401]

hello, thanks for your excellent work!

I was wondering why the new_scaling in your code is repeated (1,2), should't it be 3 rather than 6?

Scaffold-GS/scene/gaussian_model.py

Line 638 in b9e6220

new_scaling = torch.ones_like(candidate_anchor).repeat([1,2]).float().cuda()*cur_size # *0.05

thanks again and look forward to your reply!

[tongji-rkr]

#58

[ljh12401]

thanks a lot!