LLM Datasets

100,000 Instruction-Following Evaluation SFT for Chinese LLM Text Data. Between 50 and 400 words, with no fewer than 3 constraints in each prompt.All prompt are manually written to satisfy the diversity of coverage.

Large Language Model content safety considerations text data, about 500,000 in total, this dataset can be used for tasks such as LLM training, chatgpt

The data contains 203,029 groups Chinese question answering data between doctors and patients of different diseases.

2 million pairs of images and descriptions, the pictures cover various categories, including landscapes, animals, flowers and trees, people, cars, sports, industry, and architecture, along with an aesthetic subset. They depict the overall scene of the image, the details within the scene, and the emotions conveyed by the image. The description is provided in both English and Chinese languages.

This database is the interactive text corpus of real users on the mobile phone. The database itself has been desensitized to ensure of no private information of the user's (A and B are the codes to replace the sender and receiver, and sensitive information such as cellphone number and user name are replaced with '* * *'). This database can be used for tasks such as natural language understanding.

Multi-domain Customer Service Dialogue Text Data, 90,000 sets in total; spanning multiple domains, including telecommunications, e-commerce, and financial, lifestyle, business, education, healthcare, and entertainment; Each set of data consists of single or multi-turn conversations; this dataset can be used for tasks such as LLM training, chatgpt

700,000 sets of images and descriptions，the types of pictures include landscapes, animals, flowers and trees, people, cars, sports, industries, and buildings. Category and an aesthetic subset, each image has no less than two descriptions, each with one sentence; a small number of images have only one description, and the description languages are English and Chinese

11,000 Image & Video caption data of human action contains 10,000 images and 10,000videos of various human behaviors in different seasons and different shooting angles, including indoor scenes and outdoor scenes. The description language is English, mainly describing the gender, age, clothing, behavior description and body movements of the characters.

20,011 Image Caption Data of OCR in Natural Scenes, including Asian and European languages, a total of 14 languages, the collection environment includes shop plaques, stop signs, posters, road signs and other scenes, including a variety of shooting angles. The description language is English, which mainly describes the text arrangement, text content, color and other information.