search

qianqianwang68 / omnimotion

Apache License 2.0
2.12k stars 123 forks source link

Set up for 3080ti #15

Closed Ryandonofrio3 closed 1 year ago

Ryandonofrio3 commented 1 year ago

Hi there, see you trained on an A100. Only have access to a 3080ti. Should I not even bother? or just reduce the number of sampled points num_pts and the chunk size chunk_size?

I have a secondary question if there are any resources you can point to for getting this setup on a Windows machine. Not quite sure what I need to do exactly.

Thanks!

Ryandonofrio3 commented 1 year ago

Got training going! For anyone struggling ensure you run

`import torch

print(torch.version) print(torch.cuda.is_available()) ` and get True back. If not you may have to mess around with CUDA and CUDNN in your path variabless.