The easiest way would be to get a calibrator on it: connect the calibrator in place of the thermocouple and feed the SetPro a series of simulated temperatures throughout its range.
TBH though, every reasonably reputable controller I've come across in the last 20 years or so has had an error that falls well within the tolerance band for the thermocouple it is intended to be used with, even when the thermocouple is made with "special limits of accuracy" wire. I'd expect the SetPro to be no different.
With type K thermocouples, the biggest source of measurement uncertainty is almost always the thermocouple itself and that is where I'd concentrate my efforts.
If you really need that degree of accuracy, the best thing to do would probably be to buy several "same melt" thermocouples at the same time, thoroughly test your HT to find the optimum temperatures for the HT oven/thermocouple/controller combination you have, and change out the thermocouple regularly for one from the same batch to minimize the effect of "drift". For type K thermocouples, "drift" starts to become a consideration above about 1000 degC (1832 degF) and is a time-at-temperature thing: longer times at higher temperatures cause more drift. For Hobby use, annual changes would probably be fine. In daily use on Stainless steels, perhaps three-monthly and in daily use treating HSS for maximum hardness, perhaps monthly.
I tend to use type N thermocouples on the ovens I've built. Type N was developed largely to overcome the drift problem of type K. As far as I am aware though, the SetPro will not accept a type N input. The industrial controller I use will accept type N, but programming the ramp/soak profiles is a real pain compared to something designed to be user-friendly like the SetPro.
For checking (as distinct from calibrating), Tempilsticks or Orton cones can be used.
