Network Security Internet Technology Development Database Servers Mobile Phone Android Software Apple Software Computer Software News IT Information

In addition to Weibo, there is also WeChat

Please pay attention

WeChat public account

Shulou

ZTE's "Nebula R & D Model" unveiled: AI programming Assistant, 100 billion token

2025-03-29 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > IT Information >

Share

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

CTOnews.com October 12, October 11-13, 2023, during the China Mobile Global partner Conference, ZTE's "Nebula Research and Development Model" was unveiled to assist developers in requirements analysis, product design, programming, testing, version deployment, etc.

According to reports, the "Nebula R & D Model" supports the whitelist mechanism to effectively control the scope of use, as well as code feature value identification to effectively identify sensitive code fragments, sensitive word recognition mechanism to monitor and intercept sensitive content in real time, and background audit mechanism to fully trace security events.

ZTE said that since the launch of the "Nebula R & D Model" in April 2023, the number of daily active users has reached 12000, the code adoption rate has reached 40% and 45%, the code efficiency has increased by 30%, and the overall R & D efficiency has increased by 10%.

CTOnews.com learned from the official announcement that ZTE injected domain data, Know-How knowledge accumulation, hundreds of thousands of communications field technical documents and 100 billion token wireless / core network / cloud code corpus into the large model to conduct incremental pre-training, using a parallel training framework.

ZTE said: "the self-developed deployment scheme uses dynamic batch strategy, PagedAttention technology, combined with lossless model quantification, resulting in a substantial increase in throughput, with a single GPU (A800) reaching 1500tokens / s, and using only four GPU cards (A800) can meet the needs of more than 1,000 users. Compared with conventional deployment schemes in the industry, the throughput of a single GPU is increased by 10 + times and 20 + times respectively; combined with int4 quantization technology, the model size and video memory usage are reduced by half without a decrease in model accuracy. "

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.

Views: 0

*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.

Share To

IT Information

Wechat

© 2024 shulou.com SLNews company. All rights reserved.

12
Report