when training for face recognition are the training images needed after?
if I train with 5000 images, can I dump the images array and continue to predict?
yes, you can dump the images. the prediction is done entirely based on the trained model state.
you can save the trained model to xml/yml, (yml.gz even!), and in the next run, load it instead of re-training on the images again.
Asked: 2014-03-06 09:55:43 -0600
Seen: 227 times
Last updated: Mar 07 '14