search

Thinklab-SJTU / ThinkMatch

A research protocol for deep graph matching.
https://thinkmatch.readthedocs.io/
Other
826 stars 118 forks source link

No CUDA runtime is found, using CUDA_HOME='/usr/local/cuda' #51

Closed Chjx395 closed 2 years ago

Chjx395 commented 2 years ago

After I type "python train_eval.py", it raised an error "No CUDA runtime is found, using CUDA_HOME='/usr/local/cuda". Does anyone meets the same problem?

Chjx395 commented 2 years ago

After I type "python train_eval.py", it raised an error "No CUDA runtime is found, using CUDA_HOME='/usr/local/cuda". Does anyone meets the same problem?

It turns out that you should download Nvidia Docker and print "--runtime=nvidia" in the "docker run" sentence to use this docker image. Otherwise the container you open just has no gpu driver.

Forgive my silly question, since this is the 1st time that I use Docker to run a project. O(∩_∩)O

And thanks the ThinkMatch Team for your kindness and careful introduction of this project.

rogerwwww commented 2 years ago

No worries. Glad to see you resolving the problem by yourself😃

pronaykp commented 1 year ago

Bro, How can I Dockerize, would you please give the instruction! Thank you in advance.

rogerwwww commented 1 year ago

@pronaykp Do you have docker and nvidia-docker installed on your computer? If not, please follow the official guide: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker