0 devlogs
0s
Lumen is an AI coding model fine-tuned on Qwen2.5-Coder-7B-Instruct using LoRA, built for agentic coding workflows. When paired with LocalCode, it can run shell commands, use git, hunt bugs, and open pull requests — all locally, with no API keys o…
Lumen is an AI coding model fine-tuned on Qwen2.5-Coder-7B-Instruct using LoRA, built for agentic coding workflows. When paired with LocalCode, it can run shell commands, use git, hunt bugs, and open pull requests — all locally, with no API keys or cloud required. It’s available on Ollama so anyone can pull and run it in minutes.
This project uses AI
Fine-tuned Qwen2.5-Coder-7B-Instruct using LoRA as the base model. Used Claude for debugging and writing the project landing page.