“We have huge datasets that we train the models on and these are geotagged images from all over.

Report