Country211 is a dataset released by OpenAI, designed to assess the geolocation capability of visual representations. It filters the YFCC100m dataset (Thomee et al., 2016) to find 211 countries (defined as having an ISO-3166 country code) that have at least 300 photos with GPS coordinates. OpenAI built a balanced dataset with 211 categories, by sampling 200 photos for training and 100 photos for testing, for each country.
Variants: Country211
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Image Clustering | TURTLE (CLIP + DINOv2) | Let Go of Your Labels … | 2024-06-11 |
| Zero-Shot Image Classification | OpenClip H/14 (34B)(Laion2B) | Reproducible scaling laws for contrastive … | 2022-12-14 |
Recent papers with results on this dataset: