I'm trying to optimize my data pipeline according to TensorFlow's performance guide. They suggest to set the number of parallel calls to the number of threads available on the CPU.
What kind of CPUs are attached to a GPU job?
You can find the specs of the instances here: https://www.floydhub.com/features#gpu.If you want more info, you can run floyd run <instance-parameter> --follow 'cat /proc/cpuinfo'.
floyd run <instance-parameter> --follow 'cat /proc/cpuinfo'
Hope that helps.