Chinese IT and e-commerce behemoth Alibaba has unveiled two open-sourced AI models, Qwen-7B and Qwen-7B-Chat.
According to a press release, the Chinese tech and e-commerce behemoth Alibaba Group announced the availability of two open-sourced artificial intelligence (AI) models on August 3.
Qwen-7B and Qwen-7B-Chat are the names of its two large language models (LLMs), each with 7 billion parameters. According to Alibaba, these two models are scaled-down versions of the Tongyi Qiawen that the business unveiled in April.
The new models are intended to aid in integrating AI into small and medium-sized corporate operations. The company said that Qwen-7B and Qwen-7B-Chat offer several features that businesses would find compelling, including the ability to “code, model weights, and documentation will be freely accessible to academics, researchers, and commercial institutions worldwide.”
The most recent LLMs from Alibaba are also the first open-sourced products made by a Chinese digital corporation. However, it stated that organizations with over 100 million active users monthly will require a license.
The business also unveiled an update to its AnalyticDB data warehousing solution on August 1 as a vector engine, enabling its clients to quickly develop original generative AI applications.
This development came after Microsoft and Meta, on July 16 jointly published their open-sourced LLM, Llama 2. According to Meta, its Llama 2 can process twice as much context as its predecessor and can be trained using 40% more public data. Additionally open-sourced, Llama 2’s largest version has 70 billion parameters.
It necessitates a license from businesses with over 700 million monthly users, much like Alibaba’s most recent model. After deploying a Llama 2 solution to help companies create AI-powered software and solutions, Alibaba unveiled “the first training and deployment solution for the entire Llama2 series in China” on July 26.
