Sign up to take part
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
Added on March 28, 2022 12:51PM
Likes: 0
Replies: 2
Hi,
Is it possible to run code recipes with GPU? I am able to train object detection or classification tasks on the GPU. However it seems that when I run custom code recipes or notebooks, the GPU isn't used.
Hi @bkos_vis
If you are talking about R or Python recipes, the answer to your question is based truly on the code you are running. DSS doesn't change anything in the code execution and runs it "as is". In other words, if the code supposed to use GPU, it should use it.
Why are thinking GPU is not used? Are you using "nvidia-smi -l" to verify?
You are right, there was a compatibility issue between my CUDA and TensorFlow version that didn't show up in the logs. Therefore I wanted to know if DSS shielded off some resources during execution.
Thank you for your answer!