Alibaba Cloud will Support the Users of China Meta AI 'Llama'
Alibaba Cloud revealed that it will support Meta's Llama2 AI language model to businesses in China.
Alibaba's (BABA) cloud computing arm has become the first Chinese business to adopt META's open-source artificial intelligence (AI) model Llama, which gives its Chinese commercial users the right to build software separately from the model.
Main Points
- Alibaba Cloud revealed that it will support Meta's Llama2 AI language model to businesses in China.
- The move makes Alibaba the first company to present Meta's AI model in China.
- Microsoft and META are expanding their long-standing collaboration, with Microsoft being the primary partner for Llama 2..
- The news was followed by a surprise jump in Alibaba's ADRs and shares of Meta in early trade on Wednesday.
Meta just rolled out Llama2, a commercial version of Llama that aims to give businesses a low-cost alternative to the expensive proprietary models provided by OpenAI and Google. At the time, Meta stated that its preferred partner for Llama2 was Microsoft, but that it would be achieved through other partners as well.
Earlier this month Meta published a commercial version of its Llama AI model, called Llama2, which it said would offer organizations a low-cost alternative to more expensive proprietary AI models by competitors.
Alibaba Cloud has launched a training and deployment solution for the Llama2 series in China. The move enables developers to build customized large models on Alibaba Cloud. The collaboration with Meta could strengthen Alibaba's position in the cloud computing industry and help it maintain a competitive advantage in the face of increased competition and its proposed stock market listing.
"Today, Alibaba Cloud published the first training and deployment solution for the complete Llama2 series in China, inviting all developers to build customized large models on Alibaba Cloud"
The collaboration with Meta could secure a strong customer base for Alibaba's cloud business at a time when it faces tough competition domestically and looks to be listed on the stock market.
The United States is actively looking to keep many American-made technologies related to AI out of the hands of Chinese companies, particularly in the area of AI semiconductors. The Llama2 move will give Alibaba an opportunity to advance its AI hopes by staying abreast of the latest advances in technology.
For Meta, whose Facebook social media platform has been banned in China for years along with other Western platforms, it could establish stronger links with the world's second-largest economy. China is trying to compete with America in the field of AI. The country is encouraging domestic companies to build their own AI models that can challenge AI models created by American companies. Other Chinese technology giants, including Alibaba and Tencent, are actively working on their own AI models.
Like the models driving famous chatbots like OpenAI's ChatGPT and Google's Bard, Llama2 is a large machine learning model that has been trained on lots of data to produce coherent and natural-looking outputs.
According to Meta, companies with fewer than 700 million monthly active users will be allowed to use Llama2 for free. Programs with more users will need to be licensed from META.
Earlier this month, Alibaba also revealed its AI tool Tongyi Wanxiang, which it said can generate images from text prompts in English or Mandarin. It is currently available for enterprise users in beta testing, and Alibaba proposes to provide access to the tool to institutions and consumers in the coming months.
Alibaba also pointed out that if any user wants to use Llama2 to provide facilities to Chinese people, it must comply with Chinese laws and avoid process and content that harm the country.
Meta and Microsoft present the latest version of Llama
Recent achievements in AI, and especially generative AI, have captured the imagination of the general public and shown them what the creators of these technologies have long understood – their potential for helping humans with impossible tasks, economic and social Has the potential to usher in a new era of possibilities and provide new means for individuals, creators and businesses to express themselves and engage with humanity.
Meta believes that an independent approach is appropriate for building AI models today, especially in the generics area where technology is progressing quickly. By making AI models publicly accessible they can benefit everyone. Providing businesses, startups, entrepreneurs and researchers with access to advanced tools at a level that would be difficult to create themselves, driven by computing talent they may not be able to use elsewhere, to test them, in a fun way A world of possibilities will be open to you to innovate and ultimately reap benefits economically and socially.
Providing access to today's AI models means that a generation of developers and researchers can evaluate them as a community, find problems, and solve them quickly. By seeing how these tools are used by others, our own teams can learn from them, improve those tools, and plug gaps.
META has placed exploratory research, open source, and collaboration with academic and industry partners at the heart of our AI efforts for more than a decade. We have seen firsthand how open innovation creates technologies that benefit more people.
Several major language models have already been published and progress is being made by developers and researchers. These are being used by businesses as key ingredients for new generative AI-controlled experiences. Meta has been struck by the significant demand for Llama 1 from researchers - with over 100,000 requests - for access to large language models - and building on Llama 1 we've achieved some amazing results.
Now Meta is eager to open source the next version of Llama 2 and make it freely available for research and commercial use. We're also including model weights and initialization code for the pre-trained model and the conversational fine-tuning versions.
As Satya Nadella announced on stage at Microsoft Inspire, we are taking our collaboration with Microsoft to the next level for Llama 2 and expanding our efforts in generative AI. Starting today, Llama 2 is available in the Azure AI Model Catalog, enabling developers using Microsoft Azure to take advantage of cloud-native calligraphy for front-to-end, content curation protection features.
It is also optimized to run natively on Windows, giving developers a smooth workflow as they create common AI experiences for customers across multiple platforms. Llama 2 is also accessible through Amazon Web Services (AWS), Grin, and other providers.
People and businesses have benefited from the longstanding collaboration between Microsoft and META. META also offers an open ecosystem for interchangeable AI frameworks, and META co-authors research papers to advance the state of the art in AI.
META has collaborated to host PyTorch on Azure - today's leading AI framework created by the META and AI community, and META is one of the founding members of the PyTorch Foundation. Microsoft and Meta recently joined a group of advocates supporting cooperative AI models for collective action in the creation and sharing of synthetic media. META's collaboration extends beyond AI and into the metaverse to deliver enhanced experiences for the future of work and entertainment.
Now, with this expanded partnership, Microsoft and META are supporting an open approach to providing increased access to foundational AI technologies of interest to businesses worldwide. It's not just META and Microsoft who believe in normalizing access to today's AI models.
META has a diverse group of fans around the world who agree with this vision – including companies that have received META's initial response and been inspired to create innovative products with Llama 2, cloud providers who will incorporate Llama 2 into their services, research institutions who are working with META on ensuring the security and accountability of large-scale generator models, and members of the technology, academia, and policy sectors who like META Only feel the benefits.
Meta Finally Said This
Throughout our company's history, we've learned the benefits of an open source approach while innovating in other branches of the business. Our engineers created and published frameworks that are now industry standards – such as React, a leading framework for web and mobile application creation, and PyTorch, now the leading framework for AI. These became the commonly used basic framework for the entire technology industry. We believe that freely sharing today's large language models will also aid in the spread of helpful and safe generative AI.We would love to know how the version of Sansar with Llama 2 fares.