From 2f1999b2e9a08660a3fc24a2cd9bd0b6c2ac3a04 Mon Sep 17 00:00:00 2001 From: Frank Date: Sun, 16 Aug 2026 14:38:37 -0700 Subject: [PATCH] =?UTF-8?q?docs:=20frank-default-small=20live=20config=20?= =?UTF-8?q?=E2=80=94=20turbo,=20ASR,=20131072=20ctx=20verified?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- docs/FastFlowLM NPU Inference.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/FastFlowLM NPU Inference.md b/docs/FastFlowLM NPU Inference.md index eae1876..3bec0e3 100644 --- a/docs/FastFlowLM NPU Inference.md +++ b/docs/FastFlowLM NPU Inference.md @@ -199,6 +199,8 @@ Notes: `frank-default-small` now runs the NPU endpoint on Dankclaw (was the Dankserver llama.cpp service). Unit: `scripts/frank-default-small.service` (symlinked into `/etc/systemd/system`), wrapper `scripts/serve-flm-small.sh`, tunables `scripts/frank-default-small.env`. `LimitMEMLOCK=infinity` is set in the unit. Default model: `gemma4-it:e4b`. Swap `FLM_MODEL` in the env to switch. +Live config (2026-08-16): `FLM_PMODE=turbo`, `FLM_ASR=1` (loads Whisper-V3-Turbo-NPU2; exposes `/v1/audio/transcriptions`), `FLM_CTX_LEN=131072` (verified max — Gemma4 natively supports 128K and the NPU budget fits it; KV occupancy measured at 0.11% for a 123-token exchange, decode 12.0 t/s, 13 GiB headroom). + The PP split was renamed `frank-pp-head` → `frank-default-large` (2026-08-16, commit `48142f5`). ## Parallel Streams (measured 2026-08-16)