search

DecaYale / RNNPose

RNNPose: Recurrent 6-DoF Object Pose Refinement with Robust Correspondence Field Estimation and Pose Optimization, CVPR 2022
Apache License 2.0
151 stars 16 forks source link

Compatibility with RTX3090 #24

Open tensarflow opened 11 months ago

tensarflow commented 11 months ago

Hello, Thanks for sharing your work!

I am trying to reproduce your results on my local workstation. Unfortunately there seem to be serious compatibility issues with RTX3090. Specifically with Driver <-> CUDA <-> PyTorch versions. I have dangled long enough with different docker images, cuda versions etc. I came to the conclusion, that its just not going to work. I now want to implement/adapt the code to pytorch-cuda=11.7. Before I begin doing that I just want to make sure and ask you if you would recommend this. Is there any better way?

Thanks!

DecaYale commented 11 months ago

Does our provided docker work on your workstation? Of course, manually adapting the code should also be feasible.