search

vt-vl-lab / 3d-photo-inpainting

[CVPR 2020] 3D Photography using Context-aware Layered Depth Inpainting
https://shihmengli.github.io/3D-Photo-Inpainting/
Other
6.91k stars 1.11k forks source link

Requirements for GPU? #96

Closed GitHuhl closed 4 years ago

GitHuhl commented 4 years ago

Fantastic work I got this running in an Ubuntu VM. I’d like to try a dedicated bare metal installation for GPU processing. Is tensowflow required or do I just need to have the nvidia drivers installed? My card isn’t very new (compute capability 3.0 I believe) and I’m not gonna go through the pain of compiling tensorflow. Please let me know what the requirements are for GPU before I delve into installing Ubuntu and this all up again. Again great work

geborskimateusz commented 4 years ago

Hi. I would like to get more info also. I am running Ubnutu and I cannot setup project with GPU. ALways getting Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

Any hints here?

PiNaCl commented 4 years ago

If you get this answer its because you probably dont have GPU enabled. You need to have the cuda toolkit installed Have you tried following the steps in the readme with the conda environment ?

Anyway, 2 Go seems not enough.