David Amat Olóndriz, Ponç Palau Puigdevall, Adrià Salvador Palau
In this paper we introduce the FooDI-ML dataset. This dataset contains over 1.5M unique images and over 9.5M store names, product names descriptions, and collection sections gathered from the Glovo application. The data made available corresponds to food, drinks and groceries products from 37 countries in Europe, the Middle East, Africa and Latin America. The dataset comprehends 33 languages, including 870K samples of languages of countries from Eastern Europe and Western Asia such as Ukrainian and Kazakh, which have been so far underrepresented in publicly available visio-linguistic datasets. The dataset also includes widely spoken languages such as Spanish and English. To assist further research, we include benchmarks over two tasks: text-image retrieval and conditional image generation.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Image Retrieval | FooDI-ML (Spain) | A-R@1 | 0.93 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Spain) | A-R@10 | 5.8 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Spain) | A-R@5 | 3.33 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Spain) | Re-R@1 | 0.73 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Spain) | Re-R@10 | 5.67 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Spain) | Re-R@5 | 2.93 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Global) | A-R@1 | 0.005 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Global) | A-R@10 | 0.05 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Global) | A-R@5 | 0.02 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Global) | Re-R@1 | 0.01 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Global) | Re-R@10 | 0.045 | ADAPT-I2T |
| Image Retrieval | FooDI-ML (Global) | Re-R@5 | 0.03 | ADAPT-I2T |