On Thursday, Alibaba’s cloud computing arm opened sources for its two AI models based on the company’s large language model Tongyi Qianwen, becoming the first among Chinese tech giants to do so. This move is expected to further intensify global competition around open-source large models. The two public-sharing models, Qwen-7B and Qwen-7B-Chat are both miniaturized versions of Tongyi Qianwen, which was introduced in April. Alibaba Cloud’s CTO, Zhou Jingren, stated that the aim is to “enable more developers to reap the benefits of generative AI” and to support the thriving of open-source communities. The two generative AI models, equipped with 7 billion parameters each, are now freely accessible to scholars, researchers, and commercial organizations with less than 100 million monthly active users worldwide. They can be accessed through Alibaba’s AI model community and the US-based collaborative AI platform, Hugging Face. [TechNode reporting]
Related
