I am using DGL in databricks. In order to serve a model Ineed to specify the correct DGL version otherwise it will install the cpu version.
Since the version of dgl to install depends on the version of cuda, I am having a hard time figuring out a way to do this.
Is there a way to install the appropriate version of dgl via requirements.txt?
I think you can do
dgl @ https://data.dgl.ai/wheels/torch-2.4/cu121/repo.html
Or use --find-links format.
Or whatever link that matches your version requirements from this page: Deep Graph Library
The --find-links format worked fine, ty 