docs(chatterbox-fast): Phase 3 deployed — A6000, measured 5.34 GB fp32

Deployed on irv-ml1 beside live chatterbox (:8196): healthy on :8197, TTFB ~0.5s,
no starvation. Measured VRAM 5.34 GB (fp32) settles the placement: the 3090's
~3.8 GB free does NOT fit, A6000 (device 1) is the only viable card.

stacks/chatterbox-fast/.env.example

@@ -22,10 +22,10 @@ CBF_BIND=0.0.0.0
 
 # ── runtime / GPU ────────────────────────────────────────────────────
 # Device visible inside the container.
-#   1 = RTX A6000 (~12 GB free; the safe default).
-#   0 = RTX 3090  — TIGHT: turbo loads FP32 (not fp16), and the 3090 idles
-#       ~20.5 GB used (shared dev stack), leaving ~3.8 GB free. Measure the
-#       actual footprint before pinning here; it likely will NOT fit in fp32.
+#   1 = RTX A6000 (the only viable placement; ~7 GB free after this stack).
+#   0 = RTX 3090  — DOES NOT FIT: measured footprint is 5.34 GB (turbo loads
+#       FP32, not the fp16 old notes assumed), and the 3090 idles ~20.5 GB used
+#       (shared dev stack) leaving only ~3.8 GB free. Don't pin device 0.
 CBF_GPU_DEVICES=1
 
 # Default reference voice (a *.wav stem in CBF_REFERENCE_DIR, or an absolute

stacks/chatterbox-fast/README.md

@@ -52,8 +52,11 @@ docker compose build && docker compose up -d
 
 `Dockerfile` is `FROM local/chatterbox:v1` (the sibling stack's image — must exist
 on irv-ml1) + `COPY scheduler.py app.py`. GPU pin and voices/cache paths come from
-`.env` (see `.env.example`). Default GPU is **device 1 (A6000)** — turbo loads fp32,
-so the 3090's tight free VRAM likely won't fit; measure before pinning device 0.
+`.env` (see `.env.example`). GPU is **device 1 (A6000)** — measured footprint is
+**5.34 GB** (turbo loads fp32), so the 3090's ~3.8 GB free does **not** fit it.
+
+**Deployed 2026-06-02** alongside the live `chatterbox` (:8196): healthy on
+:8197, TTFB ~0.5s, no starvation, ~7 GB free left on the A6000.
 
 ## API