Haoyu Guo

icon indicating an email [email protected]

icon company State Key Lab of CAD&CG, ZJU icon location Hangzhou, China icon location Fourth year (2020-) Ph.D. student in Computer Science at Zhejiang University.

Results 6 comments of Haoyu Guo

为什么我运行strain_net.py代码的时候GPU内存极其容易爆满？

我们是用单张12G显卡训练的。你的pytorch版本是多少？建议使用1.6，参考：#8

ModuleNotFoundError: No module named 'pytorch3d.implicitron_trainer'

Hi @keloee I met the same problem, have you solved it?

Facing issue while training.

Hi, Thanks for your interest! I'm not sure what the reason is. Do you use the conda envrionment file we provide? And is this bug reproducible?

Questions about the DTU dataset.

Hi! @qihangGH I met the same problem, test views are of much lower PSNR than training views. Have you solved it?

Questions about the DTU dataset.

> @ghy0324 No, I haven't. @qihangGH @apchenstu I found that it works after modifying `fea2denseAct` in config file from `softplus` to `relu`, but I haven't figured out the reason.

Is camera pose information of the dataset provided?

> Thank you for your interest in our dataset! Does #37 answer your question (baseline between the cameras), or are you looking for the XYZ and rotation of camera for...