I want to know how can we measure the run time power when a machine learning model is running on the board (during inference)? I am using RFSOC 2x2 Academia board.
I am not exactly sure if it is available for rfsoc 2x2 (i suppose it does), you can use pmbus library to measure the power consumption. (you can look into this)
Do I have to write my entire ML model inside the with statement of pmbus datarecorder?
Sorry, I am still a noob at these stuff.
No, you can run the pmbus code in a different notebook/Python script.