five

dataDatasic/historical-culinary-archive-multimodal

收藏
Hugging Face2026-04-09 更新2026-04-12 收录
下载链接:
https://hf-mirror.com/datasets/dataDatasic/historical-culinary-archive-multimodal
下载链接
链接失效反馈
官方服务:
资源简介:
--- license: other license_name: zion-vintage-media-proprietary license_link: LICENSE task_categories: - text-generation - image-to-text - visual-question-answering language: - en tags: - historical - recipes - multimodal - ocr - alternative-data - Text-Generation - Historical size_categories: - n<1K pretty_name: Zion Vintage Recipes (Sample) --- # Zion Vintage Media: Historical Multimodal Culinary Archive (Sample) ## 📌 Commercial Licensing & Full Access **This repository contains a 50-item evaluation sample.** The full, production-ready dataset contains **10,000+ multimodal pairs** (High-Resolution Scans + Structured JSON) of unpublished, offline historical recipes spanning the 1940s to the 1990s. To purchase a commercial license for the full SQL database and high-resolution TIF/JPG scans for AI/ML training, please contact: * **Email:** admin@zionvintagemedia.com * **Website:** [zionvintagemedia.com]Vintage store(https://zionvintagemedia.com) * **Showcase:** [Dataset Explorer](https://ohcc-dataset-explore.vercel.app/) --- ## 📊 Dataset Overview This dataset provides rare, Alternative Data for training Large Language Models (LLMs) and Vision-Language Models (VLMs). Because these physical, regional church cookbooks have never been published on the public internet, they provide completely novel tokens, historical dialects, archaic measurements, and procedural logic free from standard web-scraping contamination. ### Ground Truth Quality All text has been transcribed and structured from original physical media using high-resolution CZUR archival scanners, allowing for perfect Image-to-Text grounding. --- ## 🗄️ Data Structure (22 Highly-Enriched Columns) This sample `metadata.csv` demonstrates our rigorous data engineering. It includes 22 structured columns: **Recipe Data:** * `recipe_id`, `contributor`, `contributor_role` * `yield_amount`, `oven_temperature` * `ingredients` (Serialized JSON) * `instructions` & `instruction_steps` (Serialized JSON) * `special_notes` & `data_quirks` (Captures historical oddities and OCR edge-cases) **Historical Cookbook Context:** * `book_title`, `book_author`, `publication_year` * `location`, `sales_purpose`, `historical_notes` **File Mapping:** * `source_filename`, `source_filenames` (For multimodal image pairing) --- *© 2026 Bot and Dot Retail Management LLC / Zion Vintage Media.*
提供机构:
dataDatasic
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作