CodeLlama-13b-hf Overview

CodeLlama-13b-hf is a 13 billion parameter base model within the Code Llama collection, developed by Meta. This model is part of a family of pretrained and fine-tuned generative text models, ranging from 7 billion to 70 billion parameters, all built on an optimized transformer architecture. It is specifically designed for general code synthesis and understanding, making it a versatile tool for developers.

Key Capabilities

• Code Completion: Generates relevant code snippets to complete partial code.
• Infilling: Fills in missing sections of code based on surrounding context.
• General Code Understanding: Processes and interprets code structures and logic.

Intended Use Cases

This base model is suitable for commercial and research applications in English and various programming languages. It can be adapted for a wide array of code synthesis and understanding tasks. Meta also provides specialized variants, including Code Llama - Python for Python-specific tasks and Code Llama - Instruct for instruction following and safer deployment in code assistant applications. The model was trained between January and July 2023, utilizing the same data as Llama 2 with different weights, and is governed by a custom commercial license from Meta.