zwq

Results 3 comments of zwq

> I faced the same issue. Help please! maybe you can convert the obj to off manually by above code ``` `def obj2off(objpath, offpath): ''' 将obj文件转换为off文件 :param objpath: .obj文件的路径 :param...

I also find these bugs, do you test the final result? I modified as you and used it for other task, but I can't got a good result.

感谢回答!! 有个问题还想请教一下, 我在点云压缩实验中使用这种gain增益的方式,同样是作用在bottleneck的隐变量上,但是有个问题,就是和每个lambda系数单独训练相比,加上gain增益后联合训练的结果中不同系数下的bpp的变化范围明显变窄了许多,psnr的值也十分接近。 我看你的结果好像没有这种问题,你有碰到过这个问题吗?或者你知道可能的原因在那吗?