[OC] Ai lernt, Orte auf der ganzen Welt anhand eines einzigen Bildes zu geolokalisieren

The model is a 152 layer ResNet CNN with a transformer head, organised in an ensemble architecture to process both text and image input. Images data was sourced from GoogleStreetView for training

If you are interested in learning more, I have put together a full video [explaining the model here:](https://youtu.be/coXvaZB0Sx0?si=IU1WGGZDchVc0vec)