Meta launches a community-licensed AI coding tool to the public
Under the Llama2 community license agreement, Code Llama is available for both personal and commercial use. On August 24, Meta AI announced the release of “Code Llama,” a community-licensed artificial intelligence (AI) coding tool based on the Llama2 large language model (LLM).
The new tool is a fine-tuned version of Llama2 that has been particularly trained to generate and discuss computer code, Code Llama is divided into various models, with one model optimized for general coding in a variety of languages (including Python, C++, Java, PHP, Typescript, C#, Bash, and more). Code Llama Python and Code Llama Instruct are two other models. The former is tailored to Python applications.
“Because Python is the most benchmarked language for code generation — and Python and PyTorch play an important role in the AI community — we believe a specialized model provides additional utility.”
The emphasis in these first two model versions, according to Meta, is on understanding, describing, and debating code. Code Llama Instruct, on the other hand, is the refined version of Code Llama that Meta recommends for creating code. According to the blog post, it was designed to generate “helpful and safe answers in natural language.” The models are also available in a variety of parameter sizes, allowing them to work in a variety of situations. Code Llama is available in three parameter sizes: 7 billion, 14 billion, and 34 billion.
Code Llama is accessible under the same community license agreement as Llama2, which means it can be used for personal or commercial purposes with proper attribution. This might be a huge advantage for firms and individuals with a high-need use case for LLM models for coding purposes, like finance institutions, which have historically been ignored by the AI and Big Tech community. Web3 inventors, trading bot developers, and cryptocurrency exchanges all operate in a continuously changing environment that has seen relatively little in the way of specialized B2B or B2C solutions from Big Tech for day-to-day crypto and blockchain coding cdifficulties.
Dedicated coding tools, such as GitHub’s Co-Pilot (developed with ChatGPT technology), can help developers in these underserved areas, but the fees of use can be expensive for some users, and the dearth of open-source solutions can cause issues for commercial software developers. The availability of a free, community-licensed alternative built on Meta’s highly regarded Llama2 LLM could help level the playing field for blockchain and crypto enterprises with tiny development teams.
Buy and sell crypto in minutes with 0.20% trading fees at Bitdenex Exchange