I am tasked with evaluating the performance of two separate SDK solutions. I would like to ask what's the suggested way to evaluate which one is better in terms of precision?
thanks
Asked
Active
Viewed 66 times
0
-
You need a set of data for learning and a set of data for the evaluation itself. An item within a set of data is a photo together with an identifier of the person that is shown on the photo. – mm759 Nov 16 '16 at 07:16
-
the library already comes with the trained model so we probably don't need data for learning. So I'll need to prepare data for the evaluation and feed it to both library solution and compare them, correct? – Leo Chen Nov 16 '16 at 07:28
-
You might have better luck asking on http://datascience.stackexchange.com/ , which is about machine learning and related technologies. – Eugene Mayevski 'Callback Nov 16 '16 at 07:37