price-tag-extraction

# Price tag extraction dataset This dataset contains images of price tags extracted from the Open Prices dataset, along with information extracted from this dataset. It is intended to be used to train large visual language models (LVLMs) to extract information from price tag images, as part of the [Open Prices project](https://prices.openfoodfacts.org/). For more information about the format of this dataset, please refer to the documentation on [`llm-image-extraction` datasets](https://openfoodfacts.github.io/robotoff/references/datasets/llm-image-extraction). ## Dataset creation A detailed report of the dataset creation process can be found in [this report](https://github.com/openfoodfacts/openfoodfacts-ai/blob/develop/docs/reports/2025-12-price-tag-extraction.md). We briefly summarize the steps here. ~50,000 price tag images were extracted from the Open Prices dataset, with a focus on the distribution of the countries represented in the dataset. We ensured no price tag image was also part of the [price-tag-extraction benchmark v2.0](https://github.com/openfoodfacts/openfoodfacts-ai/blob/llm-evals-price-tags-ds-v2.0/llm-evals/llm_evals/tasks/prices/price_tag_extraction/dataset.yaml). We then used Google Gemini 3 Flash with thinking configuration set to `MINIMAL` to extract the relevant information from the price tag images. ## Versions - **v1.0**: Initial version of the dataset, created in December 2025. - **v1.1**: Updated version of the dataset, with `origins` and `category` fields fixed. See the report for additional details. ## Acknowledgements This project was funded by the NLnet Foundation as part of the [NGI0 Commons Funds](https://nlnet.nl/project/OpenPrices/). Many thanks to them for their support!