Facebook launches code-generating AI
Meta Platforms (formerly Facebook) has released Code Llama, which it describes as a “state-of-the-art” large language model (LLM) capable of generating code from text prompts.
The free-to-use artificial intelligence (AI) is built on Meta’s Llama 2 LLM and is available in three models.
“Today, we are releasing Code Llama, an LLM that can use text prompts to generate code,” said Meta.
“Code Llama is state-of-the-art for publicly available LLMs on code tasks, and has the potential to make workflows faster and more efficient for current developers and lower the barrier to entry for people who are learning to code.”
The available models are as follows:
- Code Llama — the foundational core model;
- Code Llama: Python — specialised for the Python programming language; and,
- Code Llama: Instruct — a version fine-tuned for understanding natural language instructions.
The company says it built Code Llama by further training Llama 2 on code datasets.
“Essentially, Code Llama features enhanced coding capabilities, built on top of Llama 2. It can generate code, and natural language about code, from both code and natural language prompts,” said Meta.
It can also be used to complete and debug code and supports various popular coding languages, including C++, Java, PHP, Typescript, C#, and Bash.
Meta is releasing three “sizes” of Code Llama — 7B, 13B, and 34B — with different computational and latency requirements.
“The 34B model returns the best results and allows for better coding assistance, but the smaller 7B and 13B models are faster and more suitable for tasks that require low latency, like real-time code completion,” said Meta.