Code Llama
v1.0
Code Llama tool. Recently, Meta launched an AI coding tool called Code Llama, which can help users with one-click intelligent programming, easily reduce the work pressure of programmers and improve work efficiency. Users in need are welcome to download it.
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.
In addition to Code Llama, Meta also launched Code Llama-Python, which is specifically designed for Python, and Code Llama-Instrct, which can understand natural language instructions.
In Meta's benchmarks, Code Llama outperforms state-of-the-art publicly available LLM models on programming tasks.
Meta says that each specific Code Llama is not switchable/replaceable, and the company does not recommend using base Code Llama or Code Llama-Python for natural language instructions.
Meta has released Code Llama in three sizes, with 7B, 13B and 34B parameters respectively. Each model was trained using 500B code tokens and code-related data, and the 7B and 13B base and instruction models were also trained with a fill-in-the-middle (FIM) feature, allowing them to insert code into existing code, meaning You can use it now for your Code Completion Program.
