Neue Rolle benchmark (GNAI-009 5.1/5.2): Egress-Netzwerk-Quadlet gnai-testing-bench (dokumentierte Ausnahme nur fuer Datensatz-Downloads), persistentes results/- und HF-Cache-Verzeichnis, gepinntes Image gnai-bench:lm-eval-0.4.13 (python:3.13-slim per Digest, lm-eval[api]-Pin, bewusst ohne Torch), Runner-Skript mit Network-Unit-Absicherung. Live-Funde aus dem Smoke-Test (2 AIME-Samples gegen Swift 1.5): - lokale Chat-Completions brauchen --apply_chat_template (Server wendet sein --jinja-Template an, kein lokaler Tokenizer noetig) - Request-Timeout 1800 s statt lm-eval-Default 300 s (xhigh-Thinking bei ~18 t/s) - Quadlet-Netzwerke materialisieren erst mit gestarteter Unit; Rolle und Runner starten sie explizit - Ergebnis: exact_match 1.0, Per-Sample-Logs in results/; Tokenzaehlung erfolgt nachtraglich exakt ueber die /tokenize-Endpunkte der Engine Generated by Mistral Vibe. Co-Authored-By: Mistral Vibe <vibe@mistral.ai>
LLM server ghostnet-ki
This host root contains the executable infrastructure for the physical server at
10.0.0.90, split into one main directory per task area:
| Directory | Purpose |
|---|---|
host/ |
shared server baseline: Ansible control project, host roles and inventories |
llm/ |
LLM serving stack (created with the internal AI vertical slice) |
encoding/ |
CPU-based video encoding workload and its Ansible control project |
monitoring/ |
host-wide monitoring: metrics collector and dashboard |
Reusable roles that are also consumed by other host projects live under
../shared/ansible/.
The versioned architecture and operating documentation remains under
../docs/platform/llm-server/ and
mirrors the same task areas (host/, ai/, encoding/).