⚠️ This gist is archived development continues at usrbinkat/dmr
Thank you to everyone who starred, forked, commented, and reported issues here. The troubleshooting sections, model profiles, and deployment patterns in the new repository were informed by this community.
The new repository carries forward everything from this gist and adds:
- Docker Model Runner — run Qwen3.8-27B via Docker Compose with no Ollama dependency
- OpenCode integration — thin path from Docker Model Runner to local agentic coding via OpenCode
- Platform install guides — tested Linux setup (Docker Engine + DMR + NVIDIA CUDA), macOS and Windows coming
- All 12 Ollama model profiles from this gist, plus Open-WebUI integration