Key facts
Contents & fields
This dataset contains 20,846 groups of cookbook data, each group includes 4-18 images and corresponding text descriptions. Descriptions are in Chinese and English, with Chinese descriptions no less than 15 characters and English no less than 30 words. Cuisines include Chinese, Western, Korean, Japanese, etc. No field-level details are disclosed on the source page.
Research uses
Can be used for recipe recommendation, culinary education, multimodal image-text understanding, image caption generation, etc.
Information comes from the source page. Please check that page for current details and terms.
Keywords
Access and licence
Licence: License statement is still being checked | Access: check the source page
Why this is hard to get on your own
When obtaining large-scale, multi-cuisine, bilingual image-caption data for multimodal model training, data is often scattered and annotations are inconsistent.