After releasing AI models for text generation, language translation, and audio generation, Meta today released Open Source Code Llama, a machine learning system that can generate and explain code in natural language – specifically English.
Like GitHub Copilot and Amazon CodeWhisperer, as well as open source AI code engines StarCoder, StableCode and PolyCoder, Code Llama can complete and debug existing code in multiple programming languages, including Python, C++, Java, PHP . Typescript, C# and Bash.
"At Meta, we believe that AI models, but especially large language models for coding, benefit most from an open approach, both in terms of innovation and security," Meta said. in a post on her blog.
“Available public models, especially for code, can facilitate the development of new technologies that improve people's lives. By releasing code samples like Code Llama, the entire community can evaluate their capabilities, identify issues, and fix vulnerabilities.”
Code Llama, which is available in several flavors, including a version optimized for Python and a version optimized for understanding instructions (eg “Write me a function that output the fibonacci sequence”), is based on the model text generator Llama 2 that Meta released as open source earlier this month.
While Llama 2 could generate code, it wasn't necessarily good code, and certainly couldn't even approach the quality that a purpose-built model like Copilot can produce.
Download it Code Llama model.
If you are interested read more at AI blog