logoJuliaGenAI

JuliaGenAI is a GitHub organization created to promote Julia as a tool for generative artificial intelligence applications, development, and research. The organization is not intended to usurp robust tools like llama.cpp, ggml, ollama, etc.

Rather, JuliaGenAI is intended to provide a platform for furthering access to these tools in Julia, as well as to motivate the development of new tools and libraries for generative AI.

You can see which language models tend to perform best on Julia code generation here.

Organization packages

JuliaGenAI is a new organization, and as such, it is still in the process of developing its first packages. However, we have several packages that represent the type of work we hope to promote. Please see the Awesome List for GenAI meets Julia Language.

Interfaces to existing tools

Model training and inference

Why Julia?

Julia is a fantastic tool for generative AI, for several reasons.

  1. Expressiveness. Julia is a high-level language with a powerful syntax that is token-cheap – that is, you get a lot of "program" per token. It's not unlikely that AI agents will be writing code on their own in a lightly monitored way, and Julia's syntax is well-suited to low-cost AI code generation.

  2. Interactiveness. Julia is a just-in-time compiled language, which means that it's fast to run, but also fast to develop in. This is important for AI development, where the ability to quickly prototype and test ideas is crucial. It has first-in-class REPL (read-eval-print loop) support, which could potentially be very important for rapid prototype development.

  3. Performance. Julia is fast. Nuff' said.

  4. Community. Julia has a fantastic community, and it's growing. This is important for the development of AI tools, as it means that there are more people to help develop and maintain the tools that we need.

  5. Ecosystem. Julia has a fantastic and growing ecosystem built by a wonderful community of developers.

How to contribute

We are always looking for new contributors. If you have an idea for a package, or if you would like to contribute to an existing package, please feel free to reach out to us. We are also looking for people to help with documentation, testing, and other non-coding tasks.

Alternatively, you might be interested in participating in the Google Summer of Code 2024 with JuliaGenAI.

Contact

You can reach out to the organizing committee:

You can reach us on the Julia Slack & Zulip at the #generative-ai channel. You can also find us at the Discord server, or the discussion forum at discourse.julialang.org.

CC BY-SA 4.0 Cameron Pfiffer. Last modified: April 15, 2024. Website built with Franklin.jl and the Julia programming language.
Credit: The logo has been designed by Cormullion