VIAME appears to be using wrong GPU

[gcutt]

VIAME Desktop, via DIVE Desktop GUI, tried to train resnet and efficientnet classifier models on an annotated image dataset. Resulted in errors from both jobs, maybe related to insufficient memory (copied below).

More concerning - it is clear that VIAME is not using the NVIDIA GPU (GPU 1), the GPU monitor reports no applications running on the GPU, but there is activity on the Intel Graphics GPU (GPU 0).

How do I specify for VIAME to use GPU 1 (the NVIDIA GPU) instead of GPU 0 for model training? And, how to check and verify which GPU is actually being used?

Thank you.

Error from log --
[--
File "X:\VIAME\lib\python3.6\site-packages\torch_init_.py", line 124, in
raise err
OSError: [WinError 1455] The paging file is too small for this operation to complete. Error loading "X:\VIAME\lib\python3.6\site-packages\torch\lib\caffe2_detectron_ops_gpu.dll" or one of its dependencies.
Traceback (most recent call last):
--]