On March 15th, it became reported that Kai-Fu Lee’s 01.AI has currently launched its legit API. Customers can abilities the Yi series AI worthy items, including yi-34b-chat-0205, yi-34b-chat-200k, and yi-vl-plus.
yi-34b-chat-0205: A model deeply optimized per the inaugurate-source edition, with a 30% expand in instruction compliance and vastly diminished mannequin response latency. Lawful for scenarios a lot like chatting, Q&A, dialogue, collaboration, and translation.
yi-34b-chat-200k: Supports processing roughly 200K characters of Chinese (equivalent to about one e book of “Harry Potter”) or English words. Lawful for understanding more than one doc contents, massive knowledge diagnosis and mining applications right thru varied domains.
yi-vl-plus: Supports excessive-dedication image enter at 1024*1024 pixels with capabilities including image Q&A, chart comprehension, OCR, and visible reasoning. Lawful for inspecting advanced charts and screenshots for responsibilities devour details recognition, extraction, comprehension, reasoning and deal of others.
Per the introduction, the Yi-34B-Chat series API has a sooner inference crawl, which no longer most effective shortens processing time but moreover maintains very wonderful mannequin efficiency. In addition, the optimized API interface vastly reduces mannequin response latency, extra enhancing client abilities when it involves smoothness and response crawl.
Moreover, the Yi Mannequin API is entirely nicely matched with OpenAI API, allowing customers to smoothly migrate by making minor code adjustments.
SEE ALSO: Kai-Fu Lee’s AI Firm “01.AI” Announces the Start Supply of the Yi-9B Mannequin
Label in presently for 5 free articles month-to-month!