Comment by fennecbutt

I do like all the comments saying it's not good at it.

Bro, it's not even been trained to do that specifically. Imagine what an image to location specific model could do.