Shulou(Shulou.com)11/24 Report--

CTOnews.com, July 6, according to news in the market, Huawei's big model "Pangu Chat" has applied for the relevant trademark and will be released on July 7, while Huawei officials said there will be no such name as "Pangu Chat".

Hu Houyi, chairman of Huawei, attended the opening ceremony of today's 2023 World artificial Intelligence Conference and gave a performance speech. Huawei will launch Pangu 3.0 at the Huawei Cloud developer Conference on July 7, he said.

According to reports, from the general model to the industry model. Basic big model, industry big model, scene model. Pangu model has been deeply cultivated in the industry for 10 years, with 400 + business scenarios.

He said that the emergence of ChatGPT at the end of last year pushed artificial intelligence to a new tuyere. Artificial intelligence will help us rewrite everything around us.

He introduced Huawei's layout in artificial intelligence, achieved architectural innovation, ecological development and co-building computing power in deep computing power, redefined computing architecture in architectural innovation, improved the performance of peer-to-peer architecture nodes by 30%, and improved the efficiency of Teng AI cluster by 10%.

In addition, Huawei has also partnered with 5700 + Kunpeng / Quentin partners and hardware partners to achieve nearly half of the innovation capabilities of domestic large models, including scene series AI hardware 100cycles, incubation / adaptation of large model 30cycles, and Kunpeng / Jingteng developers 3.8 million +.

In terms of co-building computing power, he said Huawei is already building urban computing infrastructure to help local governments build 25 Pengteng artificial intelligence computing centers.

According to CTOnews.com information, Huawei Pangu series of basic models were officially released in 2021, including NLP (natural language processing), CV (machine vision) and scientific computing models, followed by mining, pharmaceutical molecules, meteorology, waves and other industries.

According to reports, NLP is the first Chinese pre-training model with 100 billion parameters, while the CV model reaches 3 billion parameters for the first time. Pangu CV large model is the largest CV large model in the industry, the first time to achieve both discrimination and generation ability, small sample learning ability in the industry first; Pangu meteorological large model provides second weather forecast; Zidong. Taichu is the first picture, text and sound three-mode model in the world.

For the positioning of the Pangu model, the Huawei internal team established three key core design principles: first, the model should be large, which can absorb massive data; second, the network structure should be strong, which can really give play to the performance of the model; third, it should have excellent generalization ability and can really land on the work scene of various industries.

According to Huawei's cloud official website, Pangu model is composed of NLP model, CV model, multimodal model, scientific computing model and other large models. Through model generalization, it can solve the problems of AI scale and industrialization that can not be solved under the traditional AI workshop development mode, and can support a variety of natural language processing tasks, including text generation, text classification, question and answer system, and so on.

Huawei said that the Pangu NLP model, jointly developed by Huawei Cloud, Circular Intelligence and Pengcheng Lab, has the leading ability of language understanding and model generation: in the authoritative Chinese language comprehension evaluation benchmark CLUE list, Pangu NLP model ranks first in the overall ranking, classification and reading comprehension, setting a new world historical record in three lists. The total ranking score is 83.046, leading the industry in a number of sub-tasks, taking a big step forward to the human level.

Specifically, the Pangu NLP model uses the Encoder-Decoder architecture for the first time, taking into account the understanding and generation ability of the NLP model, which ensures the flexibility of the model embedded in different systems. In downstream applications, only a small number of samples and learnable parameters can complete the rapid fine-tuning and downstream adaptation of hundreds of billions of large-scale models. This model has a good performance in intelligent public opinion and intelligent marketing.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.