The large model of "Cao Zhi" data is open to public testing, which supports long text writing with many types and complex structures.

Thanks to CTOnews.com netizen Xiao Zhan for the clue delivery! CTOnews.com, July 8 (Xinhua)-- Dagan data has released a domestic self-developed "Cao Zhi" model, which supports automatic writing and multilingual translation of long texts in multiple languages, and is open to public testing.

"Cao Zhi" takes a large number of general data and LLM as the base model, and fine-tunes Prompt Learning through a large number of general task data and domain task data.

CTOnews.com learned from the outlook data that Cao Zhi can accurately complete long text writing of many types and complex structures, automatically draft various types of documents, and has the functions of automatic typesetting, intelligent error correction, text retouching, automatic generation of summaries, etc., and will also achieve multimodal content generation, such as tables, charts and pictures in long documents. Support Chinese, English, French, German, Japanese, Korean and other dozens of language writing, assist manual work to greatly improve office efficiency; in the translation of long documents, the original title, paragraphs and other content to achieve 1:1 layout restore, provide real-time translation.

Based on the intelligent writing ability of outlook AIGC, "Cao Zhi" can adapt to all kinds of material declaration business scenarios, and quickly write all kinds of standard and non-standard documents based on the existing structured data, which can be used in professional writing scenarios such as official documents, bidding, investment bank declaration documents, legal documents and so on.