tum-nlp/text2food-mmc4

This dataset is filtered version of MMC4 Multimodal-C4 core fewer-faces dataset . It contains 144 474 pair of food image url and image caption. All the code and model in the repository.