I want to train recommender model by multi GPUs.
I used this tutorial code " [
Recommending movies: retrieval with distribution strategy
](Film önermek: dağıtım stratejisiyle alma | TensorFlow Recommenders)"
,
but my 2nd GPU’s utilization is around 1~2%.(1st GPU’s utilization is around 40%)
How can I get the second GPU to work properly?