There might be a really obvious answer I'm simply missing but on the Image Recognition Tool example there are two groups of pictures to use in the training: broken phones and not broken phones. Going through the example the first formula used is: "if contains([FullPath], 'broken') then 'Broken' else 'Not Broken' endif" which looks to only flag them all as broken since it meets the first condition.
As a result all 82 pictures are labeled as a broken phone, even those that are in the nonbroken folder, and the model returns all pictures as being classified as broken. Should the formula be switched to "if contains([FullPath], 'not broken') then 'Not Broken' else 'Broken' endif" or did I not understand the example correctly.
Thank you
