TapTechNews August 5th news, ZTE's official Weibo announced today that in the recently released Guangdong Province's generative artificial intelligence service filing announcement, ZTE Nebula R & D large model successfully completed various evaluations and passed the filing.
According to the introduction, the ZTE Nebula R & D large model supports more than 30 kinds of scenarios in different stages such as demand, design, programming, and testing and many mainstream programming languages, providing developers with a one-stop, intelligent R & D experience.
The ZTE Nebula R & D large model is claimed to have code generation ability reaching the level of GPT-4, and it has significantly surpassed GPT-4 Turbo in unit test accuracy and coverage, helping to improve coding efficiency by 30% and overall R & D efficiency by 10%.
Nebula R & D large model supports a whitelist mechanism to effectively control the scope of use, as well as a code eigenvalue recognition to effectively identify sensitive code fragments, a sensitive word recognition mechanism to monitor and intercept sensitive content in real time, and a back-end audit mechanism to completely trace back security events, etc.
ZTE said that in April 2023, the use of Nebula R & D large model started. TapTechNews inquiry found that ZTE injected domain data, Know-How knowledge accumulation, hundreds of thousands of technical documents in the communication field and 100 billion tokens of wireless/core network/cloud code corpora, etc. into the large model for incremental pre-training and used a parallel training framework.