Improving DeepStack Models?

Nov 16, 2017
6
2
I am new to the AI stuff, but I wanted to clarify a question. Many of the standard and custom models work well, but I would like to teach them specifics about my video.
Is it possible to take a .pt model and add training for additional images or items to it?

It seems like you either can use the standard model or create entirely new models. I am training a custom one now, but I would rather add my information to improve upon an existing one rather than recreate the wheel. Is this only possible by having the original dataset?
 
Last edited: