In the months since ChatGPT was released, the world has gained a new understanding of the capabilities of AI.
ChatGPT started the battle, and now there is one more player!
The Ali version of ChatGPT suddenly officially announced that it is officially open to the public for enterprise invitation testing. It was developed by Dharma Institute and is called "Tongyi Qianwen".
As the name suggests, Ali officially joined the ChatGPT battle! What is the strength of this Ali version of ChatGPT "Tongyi Qianqian"?
Ali version ChatGPT record
As a large language model, Tongyi Qianwen mainly focuses on text generation, that is, it can also "ask and answer" like ChatGPT. In addition, it also has a "treasure bag" function. It's also very cool, and what everyone is generally interested in is recipe generation, rainbow fart generator and free love letter writing. It is equivalent to a toolbox, which can quickly generate various types of copywriting.
What tasks can Tongyi Qianwen accomplish, what productivity can be liberated, and how much skill can it achieve ChatGPT?
Judging from the actual performance, there is still a big gap between Tongyi Qianwen and the leading GPT 4.0, but compared with some previous large AI models, Tongyi Qianwen has already developed in terms of question-and-answer logic, information retrieval, and sentence understanding. Obvious progress, I am now full of confidence in the future of China's AI big model.
Where did Tong Yiqian come from?
The birth of Tongyi Qianwen is not accidental, but the result of years of technology accumulation.
In 2019 , Ali started the research and development of large-scale models. At that time, the large-scale language model StructBERT released by Ali surpassed Google, Microsoft, and Facebook, and reached the top of the CLUE list.
In 2021 , Alibaba Dharma Academy released the first domestic multi-modal large-scale model M6 with over 10 billion parameters, claiming that the number of neurons is 10 times that of human beings.
At the World Artificial Intelligence Conference in 2022 , Ali also released the "Tongyi" large-scale model series, and announced that the relevant core models will be open sourced to developers around the world.
Regarding the technical details of Tongyi Qianwen, the official Ali Dharma Academy did not disclose detailed information. Although the official caliber is low-key, everyone understands it, and the domestic technology circle explodes directly. Powerful players enter the arena, and the competition for large models really begins.
Just as ChatGPT was born out of OpenAI's GPT series, Baidu Wenxinyiyan was developed from Ernie's large-scale model, and Ali is also one of the first large-scale technology companies in China to start developing large-scale models.
Among them, after multiple iterations, M6 has achieved a parameter scale of 10 trillion levels, and combined with the business needs of Alipay and Taobao, M6 is the first to achieve commercialization in China. At last year's WAIC (World Artificial Intelligence Conference), Ali also released the Tongyi large-scale model series. The core models are all open source.
Everyone is a low-code developer
The potential of generative AI products is enormous, but companies that can make similar applications and commercialize them are rare. Many companies believe that AI can help the construction of the enterprise, but the problems accumulated within the enterprise over the years are very complicated. For example, the data required by an artificial intelligence platform of the enterprise needs to be obtained from dozens of systems, and the docking of these systems takes a lot of time And energy, the infrastructure of different companies belonging to the same field is not exactly the same, and simple copying and pasting cannot achieve the effect.
In the development of low-code, there is a popular saying: Everyone is a low-code developer.
In the cloud computing and software industry, low-code development and even no-code development have been popular concepts in the past two years. The core logic is to use the code library to quickly rewrite existing development samples, and the labor cost in the entire development process is close to zero.
A complete set of bases means common solution capabilities in most fields.
JNPF relies on the principle of code development technology, so it is different from the traditional pain points of long development delivery cycle, difficult secondary development, and high technical threshold. Most of the application construction is realized by dragging and dropping controls, which is simple and easy to use. By providing developers with a visual application development environment, it reduces or eliminates the need for application development to write native code, and then realizes a convenient construction of applications. The development platform quickly assists R&D personnel to quickly build a set of solutions suitable for enterprise development.
Open source link: https://www.yinmaisoft.com/?from=csdn
In the era of large models, China's power is accelerating its competition
Compared with the current industry benchmark ChatGPT, Tongyi Qianwen still has a lot of room for improvement. Ali also revealed that according to internal test feedback, this large model is being iterated rapidly.
At the same time, Ali is one of the few companies in the world that has a leading layout in terms of algorithms and computing power, and it also has natural advantages in terms of computing power. ChatGPT has been on fire until now, and the domestic demand for domestic generative large models with sufficient competitiveness is increasing day by day.
The world is changing, and a new era of AI is coming. Various uncertainties have once again highlighted the value of self-development of technology.
Fortunately, this time, our starting line is not so far apart.
The game doesn't end overnight, and now, the real race begins.
Finally, what questions do you want to ask Tongyi Qianwen? Please leave a message in bold.