omp-ninfer

OMP NInfer is a packaging of an integration between Oh My Pi, Neroued's NInfer, and Qwen3.8 27B. It resumes from saved model state, allowing long local Qwen coding sessions to continue after an inference-server restart. Qwen3.8 27B is the model, Neroued's NInfer is the inference engine, and Oh My Pi is the coding agent. The project packages their integration, explicit continuation, and durable checkpoints into exact, qualified releases.

Eligibility for v0.7.3 is limited to specific hardware. RTX 5090: documented Windows 11 + Docker Desktop/WSL2 runtime route. RTX 4090: documented native Windows 11 route. RTX 3090 is deferred for v0.7.3. v0.7.3 repins the client to OMP 18.2.3, while runtime bytes, model, serving settings and memory floors stay at v0.7.2.

Checkpoints can leave the machine. Export verified session checkpoints, keep replicas on another host or NAS, and import them back to local storage for restore. Restore needs the same runtime binary, model, profile, and session credentials. This is not arbitrary GPU-to-GPU migration or automatic failover.

The architecture is Oh My Pi (client) → OMP NInfer (qualified appliance + releases) → NInfer (engine) → Qwen3.8 27B (model). It is loopback-only, bearer-authenticated, fail-closed instead of cloud fallback, every byte hash-pinned. Community project; not affiliated with or endorsed by Oh My Pi, Qwen, or NVIDIA.