Using experiment manager on single GPU

Hi, I’m training CNNs with various architectures and hyperparameters using experiment manager. What is the intended execution mode if I train it on single GPU (NVIDIA A5500)? Is it just sequential execution mode running in GPU execution environment? Or should I experiment with parallel execution mode even though I have only single GPU? Thank you for any advice!

Answer 1

[Joss Knight]

It does depend on the balance of CPU and GPU work in your experiment, but as a general rule parallel execution will gain you nothing while making you more likely to exhaust GPU memory.