Flitto資料市集提供100多種語言與超過23個專業領域的AI學習與評估資料集，涵蓋文字、語音、圖片、影片等已授權驗證的資料。歡迎先下載免費樣本查閱，如欲購買，請隨時洽詢。
A labeled image dataset capturing surface defects on automotive body panels, including scratches, dents, and paint irregularities. Built to support vision inspection model training in automotive manufacturing processes.
A red-teaming image dataset designed for AI safety evaluation. Includes adversarial inputs, harmful content bypass attempts, and policy-violating scenarios to support model safety assessment and AI governance enhancement.
An English multi-turn conversational speech dataset covering end-to-end business negotiations across diverse B2B domains, including SOW scope alignment, cost and timeline risk discussions, and contract revision workflows between stakeholders.
A multi-turn conversational speech dataset in the Gulf Arabic dialect, covering a wide range of real-life topics such as everyday conversations, hobbies, shopping experiences, and opinion exchanges related to social media.
A multi-turn conversational speech dataset in the Egyptian Arabic dialect, covering a wide range of real-life topics such as everyday conversations, hobbies, shopping experiences, and opinion exchanges related to social media.
A multi-turn conversational speech dataset in the Levantine Arabic dialect, covering a wide range of real-life topics such as everyday conversations, hobbies, shopping experiences, and opinion exchanges related to social media.
A multi-turn conversational speech dataset in the Maghrebi Arabic dialect, covering a wide range of real-life topics such as everyday conversations, hobbies, shopping experiences, and opinion exchanges related to social media.
A multilingual OCR image dataset for menu images in seven languages, including Korean, English, Chinese, and Japanese, comprising original images, translated-text images, and inpainted images with professionally reviewed annotations.
An English handwritten document OCR image dataset with bounding-box annotations for multilingual document understanding, OCR model training, and layout-aware text recognition.
An English OCR and document parsing sample dataset for manufacturing documents, built with bounding-box JSON annotations for PoC, Document AI, and model evaluation use cases.