To listen click here
Hong Kong, Aug 4 – As generative AI race heats up in the US, the Chinese company Alibaba Cloud has announced to open source two of its large language models (LLMs) for researcher community.
The two open-source models, Qwen-7B and Qwen-7B-Chat, have each been trained on 7 billion parameters, reports the South China Morning Post.
The AI models and their codes and documentation “will be made freely accessible to scholars, researchers and commercial institutions worldwide through
Alibaba Cloud’s AI model repository ModelScope and the US collaborative AI platform Hugging Face,” the report mentioned.
Companies with fewer than 100 million monthly active users will be allowed to deploy the open-source models for commercial use free of charge, the report mentioned.
Alibaba Cloud is set to be spun off from its parent company Alibaba Group next year.
“We aim to promote inclusive technologies and enable more developers and small and medium-sized enterprises to reap the benefits of generative AI,”
Zhou Jingren, chief technology officer of Alibaba Cloud Intelligence, was quoted as saying.
Llama 2, the open-source LLM created by Meta, has been added to Alibaba’s ModelScope platform.
In May, reports surfaced that Chinese internet giant Alibaba was making significant job cuts in its Cloud unit, reportedly around 7 per cent of its workforce, as it plans separate IPOs for its various business groups.
In March, Alibaba Group planned to split into six business groups and launch separate public listings, triggering mass layoff, according to a Nikkei Asia report.
Alibaba CEO Daniel Zhang had earlier detailed the restructuring of its Cloud division.