Load model on multiple GPUS

Hi,

I tried loading the model on multiple GPUs, but am stuck as if I try to do it with pytorch the core functions stop working. Is there an easy way to simply load the model on two GPUs at the same time?

The issue mentions loading the model with PyTorch across two GPUs, but names no files, tests, or entry points. Start by locating the model-loading path and checking how PyTorch device placement is handled. Done means the model loads on two GPUs while its core tracking functions continue to work.

Written by the indexing model from the issue text.