d-ailin / GDN

Implementation code for the paper "Graph Neural Network-Based Anomaly Detection in Multivariate Time Series" (AAAI 2021)
MIT License
481 stars 141 forks source link

CUDA is not available #87

Open xxxdongjiu opened 8 months ago

xxxdongjiu commented 8 months ago

When I tested whether cuda was available on the console, no errors were reported. image However, when the "bash run.sh 1 msl" command is executed on the terminal, a message is displayed indicating that cuda is unavailable. image Is it because cuda version is inconsistent or there are other reasons, can you give me some help?

d-ailin commented 8 months ago

The script "bash run.sh 1 msl" indicates running with "cuda:1". May I know if there is more than 2 gpus on your device? If not, then you could run "bash run.sh 0 msl" instead. Feel free to let me know if there are other issues.

Changshengli0 commented 8 months ago

1709699595089I'm having the same problem as the person who asked, after using the bash run.sh 0 msl command it still gives me an error, can you help me please?

d-ailin commented 8 months ago

I guess it might be due to the mismatched pytorch version. May I know what pytorch and cuda version you are using? Or you could also refer to the readme.md to create a new virtual environment with the exact package versions. Thanks.