Closed xuxumiao777 closed 1 year ago
which version of nerfacc are you using?
@liruilong940607 nerfacc version is (0.2.2)
Could you try with the latest version pip install nerfacc==0.3.2
? That was an issue at a certain version (I don't remember which one) but I fixed it in the lastest version.
Starting from this version we will have pre-built wheels provided so that you can directly install it:
Hi, ruilong Thanks, everything works fine for me now. The problem is not in nerfacc. It is the mismatch of one of the dependency of tinycudann.
have a nice day! Xu
Ruilong Li(李瑞龙) @.***> 于 2023年2月1日周三 01:58写道:
Starting from this version we will have pre-built wheels provided so that you can directly install it:
https://github.com/KAIR-BAIR/nerfacc/releases/tag/v0.3.4
— Reply to this email directly, view it on GitHub https://github.com/KAIR-BAIR/nerfacc/issues/143#issuecomment-1411286627, or unsubscribe https://github.com/notifications/unsubscribe-auth/AZSJ2JQ3HRJT6I5LREUBVXTWVGYMLANCNFSM6AAAAAATZLUK7U . You are receiving this because you authored the thread.Message ID: @.***>
Hi, ruilong Thanks, everything works fine for me now. The problem is not in nerfacc. It is the mismatch of one of the dependency of tinycudann. have a nice day! Xu Ruilong Li(李瑞龙) @.> 于 2023年2月1日周三 01:58写道: … Starting from this version we will have pre-built wheels provided so that you can directly install it: https://github.com/KAIR-BAIR/nerfacc/releases/tag/v0.3.4 — Reply to this email directly, view it on GitHub <#143 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AZSJ2JQ3HRJT6I5LREUBVXTWVGYMLANCNFSM6AAAAAATZLUK7U . You are receiving this because you authored the thread.Message ID: @.>
Could you please tell me which dependency of tinycudann is mismatch? How did you solve the problem?Thanks a lot!
Hi, ruilong I have some trouble when training. The program stucks when setting up NerfAcc. I have tried to empty cache in torch_extensions and rerun the script but it still fails. My GPU is 3060 Laptop, python3.9, cuda 11.7. Thanks and hope for your reply.