AMDMIGraphX
AMDMIGraphX copied to clipboard
Enable MultiStream execution on DLM based models
This will help lower batch sizes, need to determine which models it benefits. Needed for exec_dashboard models