Coding with LLMs on felix

Coding with LLMs on Felix

Using and configuring Felix (felix.loria.fr) as coding assistant:

Discover how to use and configure local generative AIs (felix.loria.fr) as coding assistants.

A few weeks ago, we made available two machines (tigre and jaguar) on the local network, accessible via felix.loria.fr for using generative AIs: https://sed-nge.inria.fr/platforms/ollama/index.html

In this tuto techno, you will see the following points:

• Using AIs via Open WebUI: configuration, system prompt personalization, function usage, documents, etc.
• Local installation of Open WebUI on your machines and configuration to leverage knowledge bases (RAG).
• Using servers via the Ollama client for Python or another programming language.
• Using G5K to run heavy processes.
• Configuring coding assistants like aider and Continue.
• Developing a complete small web application with the assistant: code, tests, documentation.

Prerequisites:

• Basic programming skill
• Understanding the limitations of LLMs (text generation)

Technologies used:

• Your favorite IDE: Visual Studio (Windows) or Visual Studio Code (Linux)
• Programming language: Python or another.