Skip to content

Latest commit

 

History

4 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

qServer — local Qwen inference stack

A bulletproof local inference stack for three Qwen models on a single dual-GPU workstation (RTX 4090 + RTX 5060 Ti + system RAM), managed by llama-swap behind one OpenAI-compatible endpoint. Models are mutually exclusive — one resident at a time, loaded on demand by name.

model id (alias) placement speed
27B dense + vision (Qwen3.8) Q6_K qwen-38-27b (qwen38, qwen36, qwen36-q6) 4090 + 5060 Ti ~73 t/s code · 48 @100K
35B-A3B MoE qwen-35b (qwen36-35b) 4090 + 5060 Ti ~206 t/s
122B-A10B MoE qwen-122b (qwen35-122b) 4090 + 5060 Ti + RAM ~37–40 t/s

All: 131072 context, q8_0 KV, MTP speculative decoding, thinking on. The 27B is a native vision-language model (images + video) and replaced the retired 3.6-27B dense line (Q4 + Q6); its legacy aliases (qwen36, qwen36-q6, qwen36-text) carry forward so existing clients keep working.

Docs

  • API.md — the OpenAI-compatible API, every endpoint, and the metrics anyone building on it needs.
  • USAGE.md — day-to-day commands and gotchas.
  • RESULTS.md / LEARNINGS.md — the full build record: measurements, bugs hit, and why each decision was made.

Layout

run-text-*.sh        llama-server launch scripts (one per model, every hard-won flag encoded)
router/proxy.py      OpenAI-compatible router :8000 (tool-budget guard, /load, /unload)
swap/config.yaml     llama-swap model manager :9000 (mutual-exclusion group)
systemd/             units + boot verification
templates/           chat templates (v19)
gates*.sh, regress.sh, canary.py, needle35.py, soak*.py   regression / gate suites

Not included

Model weights (models/, *.gguf, ~125 GB) and the CUDA-built llama.cpp/ are gitignored — build llama.cpp with CUDA 12.8 and fetch the GGUFs from the unsloth repos referenced in the run scripts.

Requires CUDA 12.8 specifically (13.1 breaks MMQ, 13.2 produces gibberish on this arch).

About

No description, website, or topics provided.

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages