vh
b75f020cc9
Adds a third TTS to the irv-ml1 fleet. IndexTTS-2 is Bilibili's
emotion-controllable zero-shot TTS (paper 2506.21619). Distinguishing
capability vs the existing two: timbre and emotion are disentangled —
clone a voice's timbre from one reference and the emotion from a
different reference, OR set emotion via 8-vector, OR derive it from a
text description. Neither CosyVoice 3 nor Qwen3-TTS-1.7B-Base does
this cleanly in English.
Wrapper is owned end-to-end (~150 lines in app.py) — the only existing
FastAPI fork (csllpr/index-tts-fastapi) targets v1 and is a dormant
single-commit repo. Upstream IndexTTS-2 ships only a Gradio webui.
Layout follows the qwen3-tts pattern:
stacks/index-tts/
Dockerfile — CUDA 12.8 base, IndexTTS pinned to a SHA
app.py — FastAPI: POST /v1/audio/speech + /v1/voices
entrypoint.sh — one-time HF snapshot_download of the weights
compose.yaml — env-driven, GPU pinning support, bind mounts
.env.example — port 8192, fp16, paths
README.md — API examples + comparison vs the other TTS
playbooks/deploy-index-tts.yaml — elway playbook for irv-ml1
Voice and emotion libraries are flat host dirs of WAVs, bind-mounted.
Drop a new <name>.wav and /v1/voices picks it up immediately.
License caveat: IndexTTS-2 weights ship under a custom Bilibili
license (free at our scale, not OSI-open). README documents it.
64 lines
2.5 KiB
YAML
64 lines
2.5 KiB
YAML
# IndexTTS-2 — Bilibili's emotion-controllable zero-shot TTS, served
|
|
# behind our own thin FastAPI wrapper (stacks/index-tts/app.py).
|
|
#
|
|
# Why this stack exists alongside qwen3-tts and cosyvoice:
|
|
# IndexTTS-2 disentangles timbre from emotion — emotion can be cloned
|
|
# from a separate audio reference, set via 8-vector, or derived from
|
|
# free text. Neither qwen3-tts nor cosyvoice expose this cleanly in
|
|
# English. See stacks/index-tts/README.md for the full rationale.
|
|
#
|
|
# Build: image is local, built from the Dockerfile in this dir. Pinned
|
|
# upstream SHA lives in .env as INDEX_TTS_SHA so rebuilds are
|
|
# reproducible.
|
|
#
|
|
# Model: ~5-7 GB IndexTTS-2 weights download on first start via the
|
|
# entrypoint, persisted under ${INDEX_TTS_CACHE_DIR}.
|
|
#
|
|
# License note: weights carry a custom Bilibili license (free at our
|
|
# scale, but not OSI-open). The wrapper code is ours, MIT-by-default.
|
|
#
|
|
# All tunables live in .env — edit that, not this file.
|
|
|
|
services:
|
|
index-tts:
|
|
image: local/index-tts:${INDEX_TTS_TAG}
|
|
build:
|
|
context: .
|
|
dockerfile: Dockerfile
|
|
args:
|
|
INDEX_TTS_SHA: ${INDEX_TTS_SHA}
|
|
container_name: index-tts
|
|
restart: unless-stopped
|
|
runtime: nvidia
|
|
ports:
|
|
- "${INDEX_TTS_BIND:-0.0.0.0}:${INDEX_TTS_PORT}:8000"
|
|
environment:
|
|
- NVIDIA_VISIBLE_DEVICES=${INDEX_TTS_GPU_DEVICES:-all}
|
|
- INDEX_TTS_MODEL_DIR=/app/checkpoints
|
|
- INDEX_TTS_VOICES_DIR=/app/voices
|
|
- INDEX_TTS_EMOTIONS_DIR=/app/emotions
|
|
- INDEX_TTS_FP16=${INDEX_TTS_FP16:-1}
|
|
- INDEX_TTS_DEVICE=${INDEX_TTS_DEVICE:-}
|
|
- INDEX_TTS_LOG_LEVEL=${INDEX_TTS_LOG_LEVEL:-INFO}
|
|
volumes:
|
|
- ${INDEX_TTS_CACHE_DIR}:/app/checkpoints
|
|
- ${INDEX_TTS_VOICES_DIR}:/app/voices
|
|
- ${INDEX_TTS_EMOTIONS_DIR}:/app/emotions
|
|
healthcheck:
|
|
# Match Dockerfile's healthcheck. Compose-level entry overrides the
|
|
# image-level one if anything ever needs tweaking per-deploy.
|
|
test: ["CMD-SHELL", "wget -q -O /dev/null http://localhost:8000/healthz || exit 1"]
|
|
interval: 30s
|
|
timeout: 10s
|
|
retries: 3
|
|
# First boot: ~5-7 GB HF download + IndexTTS-2 import (touches several
|
|
# auxiliary HF repos for MaskGCT, campplus, BigVGAN, w2v-bert) + initial
|
|
# CUDA warmup. Generous deadline to ride that out.
|
|
start_period: 600s
|
|
labels:
|
|
- homepage.group=AI Systems
|
|
- homepage.name=IndexTTS-2
|
|
- homepage.icon=mdi-account-music
|
|
- homepage.description=Emotion-controllable TTS w/ voice cloning (irv-ml1)
|
|
- homepage.href=http://10.100.79.3:${INDEX_TTS_PORT}
|