vh
ca16db73e0
Profiling the index-tts deploy log (2057 lines) showed ~25% was just
pip's per-package Downloading / Collecting / Requirement-already /
progress-bar spam — useless for ops, hard to scan when something
actually breaks.
Three changes across the four TTS deploy playbooks:
1. Pulls (Kokoro): add --quiet. 6.5 GB pull no longer floods the log
with per-layer progress redraws. Final "X Pulled" still prints.
2. Builds (VibeVoice, Chatterbox, IndexTTS-2): add --progress=plain
to stop the BuildKit TUI from littering the captured log with
carriage-return overdraws, then pipe through a grep filter that
drops pip's noisy lines but keeps:
- buildkit step transitions (#NN [stage])
- DONE / CACHED / ERROR markers
- apt + build-stage messages
set -o pipefail keeps a real build failure from being swallowed
by the grep's exit code.
Net effect: ~25% smaller logs, much more scannable; full visibility
into step progress and errors preserved.
94 lines
3.4 KiB
YAML
94 lines
3.4 KiB
YAML
# Deploy Chatterbox Turbo (Resemble AI's low-latency English TTS w/
|
|
# voice cloning) via the devnen/Chatterbox-TTS-Server wrapper to
|
|
# irv-ml1.
|
|
#
|
|
# Builds the image locally from devnen's Dockerfile.gpu via docker
|
|
# buildx git URL context. ~8-10 min cold build (CUDA + torch +
|
|
# Chatterbox deps). First start pulls Chatterbox-Turbo weights (~6 GB)
|
|
# into the HF cache.
|
|
#
|
|
# Usage:
|
|
# scripts/elway irv-ml1 --playbook playbooks/deploy-chatterbox.yaml
|
|
#
|
|
# Idempotent — every step is creates-/when-gated; rerun is safe.
|
|
|
|
vars:
|
|
compose_dir: /opt/docker/compose/chatterbox
|
|
reference_dir: /worktank/chatterbox/reference_audio
|
|
cache_dir: /worktank/chatterbox/cache
|
|
host_port: "8196"
|
|
|
|
steps:
|
|
# ── host-side dirs ──────────────────────────────────────────────────
|
|
|
|
- name: Ensure /worktank/chatterbox root exists (one-time, sudo)
|
|
shell: mkdir -p /worktank/chatterbox
|
|
sudo: true
|
|
creates: /worktank/chatterbox
|
|
|
|
- name: Chown /worktank/chatterbox to lkraven
|
|
shell: chown lkraven:lkraven /worktank/chatterbox
|
|
sudo: true
|
|
when: '[ "$(stat -c %U /worktank/chatterbox)" != lkraven ]'
|
|
|
|
- name: Ensure reference-audio dir exists
|
|
shell: mkdir -p {{ reference_dir }}
|
|
creates: "{{ reference_dir }}"
|
|
|
|
- name: Ensure cache dir exists
|
|
shell: mkdir -p {{ cache_dir }}
|
|
creates: "{{ cache_dir }}"
|
|
|
|
- name: Ensure compose dir exists
|
|
shell: mkdir -p {{ compose_dir }}
|
|
creates: "{{ compose_dir }}"
|
|
|
|
# ── deploy compose files ────────────────────────────────────────────
|
|
|
|
- name: Upload compose.yaml
|
|
upload:
|
|
src: stacks/chatterbox/compose.yaml
|
|
dest: "{{ compose_dir }}/compose.yaml"
|
|
mode: "0644"
|
|
|
|
- name: Seed .env from template (only if absent)
|
|
upload:
|
|
src: stacks/chatterbox/.env.example
|
|
dest: "{{ compose_dir }}/.env"
|
|
mode: "0644"
|
|
when: "[ ! -f {{ compose_dir }}/.env ]"
|
|
|
|
# ── build + bring up ────────────────────────────────────────────────
|
|
|
|
- name: docker compose build (~8-10 min first time; cached after)
|
|
# See deploy-vibevoice.yaml for the rationale — same filter pattern.
|
|
shell: |
|
|
set -o pipefail
|
|
cd {{ compose_dir }} && docker compose build --progress=plain 2>&1 \
|
|
| grep -vE '^#[0-9]+ [0-9.]+ (Downloading|Collecting|Requirement|Using cached|Installing collected|Successfully (installed|built)|Saved /|━|Resolved|Prepared|Built)'
|
|
|
|
- name: docker compose up -d
|
|
shell: cd {{ compose_dir }} && docker compose up -d
|
|
|
|
- name: Wait for /health to respond (allow ~15 min for model download + warmup)
|
|
shell: |
|
|
for i in $(seq 1 180); do
|
|
curl -sf -o /dev/null --max-time 3 http://localhost:{{ host_port }}/health && exit 0
|
|
sleep 5
|
|
done
|
|
exit 1
|
|
changed_when: "false"
|
|
|
|
verify:
|
|
- name: /health returns 200
|
|
shell: curl -sf -o /dev/null http://localhost:{{ host_port }}/health
|
|
changed_when: "false"
|
|
|
|
- name: /v1/audio/voices returns valid JSON
|
|
shell: curl -sf http://localhost:{{ host_port }}/v1/audio/voices | grep -q 'voice\|alloy\|echo'
|
|
changed_when: "false"
|
|
|
|
- name: Container is running
|
|
shell: docker inspect chatterbox --format '{{.State.Status}}' | grep -q running
|
|
changed_when: "false"
|