Code Llama large model. Today I am sharing the Code Llama large model, which is launched by Meta. Next, leave the programming work to the Code Llama large model. You can enter your needs with one click and complete it easily and efficiently.
Meta has now launched an AI coding tool called Code Llama, which is billed as "the most advanced large-scale coding language model."
This model is based on the Llama 2 large-scale language model, which can be understood as a "code-writing-specific version of Llama 2" and can be used to "generate new code and debug human-written work". It is currently available on Github.
Meta points out that Code Llama can generate the code you want and natural language about the code based on code core natural language prompts/requirements, or refine and debug when pointed to specific code.
"Programmers are already using LLMs to assist with a variety of tasks, from writing new software to debugging existing code," Meta said in its blog. "The goal is to make developer workflows more efficient so they can focus on to the most people-centered aspects of their work.”
Meta claims that Code Llama outperforms existing publicly available LLMs in benchmark tests, but does not specify which models it was tested against.
The company said that Code Llama scored 53.7% on the HumanEval code benchmark and was able to accurately write code based on text descriptions; it scored 56.2% on MBPP, which is comparable to the state-of-the-art ChatGPT.
It is worth mentioning that GitHub launched Copliot based on GPT-4 in March this year to help users quickly write and check code, and can also rewrite old code for updates, but it has faced legal action for being accused of infringing copyright laws. .
In addition, Amazon AWS also has a CodeWhisperer, which can write, check and update code; and Google also has a tool for writing code in AlphaCode, but it has not yet been released to the public.