| task_categories: | |
| - image-to-text | |
| - text-to-image | |
| language: | |
| - ar | |
| - en | |
| - ru | |
| - it | |
| - es | |
| - ko | |
| - pl | |
| - tr | |
| - zh | |
| - de | |
| - fr | |
| - ja | |
| size_categories: | |
| - 1K<n<10K | |
| # Dataset Summary | |
| The expanded XTD-11 dataset, now including Arabic, enhances the original XTD collection. This dataset introduces a 1,000-image multi-lingual MSCOCO2014 caption to test multimodel in zeroshot image or text retrieval in 11 Languages. | |
| # Dataset |