I am working on DGX system having 4 GPU cores with docker. I was trying to access the GPU 0 to train my tensorflow code. I used
import os
os.environ[“CUDA_DEVICE_ORDER”]=“PCI_BUS_ID”
os.environ[“CUDA_VISIBLE_DEVICES”]=“0”
I could not able to access the gpu from DGX system. Anyone please help me to solve this issue.