In the docker file, CUDA 11.8 is used, while in the prepare environment documentation, CUDA is used 11.8.
Do you know which we should use? Please advise. Thanks.
Note: I set up a Conda environment using 11.7, and I'm wondering if updating CUDA to 11.8 will improve execution performance.
Any guidance/advice is much appreciated. Thanks.