CPA training set file tail

Hi folks,

I notice that every training set that I save after working with the CPA classifier has a very large line appended at the end, which causes the file to be very large relative to what it contains.

For example, one training set with 834 objects is 20K without the last line, but 5.5M with the last line. Does anyone else see this behaviour?



Hi Lee,

This added information is an encoded version associated measurements for the objects in your training set. It serves a cache in order to confirm that the per-object data hasn’t changed in the database, and if not, allows for quick loading of the data so you can find rules immediately even as the image tiles are loading.