Inspired by tile2vec, pretrains the models by training the models to make embeddings that are far away from each other more different than embeddings that are close to one another.
It's a less rigid way of communicating the latlon information to the models
Inspired by tile2vec, pretrains the models by training the models to make embeddings that are far away from each other more different than embeddings that are close to one another.
It's a less rigid way of communicating the latlon information to the models