Files
vllm-omni/examples/online_serving/text_to_speech/indextts2/run_server.sh
T
BeatSeat 044240d8eb [Model] Add IndexTTS2 text-to-speech support (#3838)
Signed-off-by: BeatSeat <wendavid552@gmail.com>
Signed-off-by: David <david@podcast.ai>
Signed-off-by: wendavid552 <wendavid552@gmail.com>
Signed-off-by: WenDavid <davidzhangcf@foxmail.com>
2026-06-22 19:24:01 +00:00

26 lines
782 B
Bash

#!/bin/bash
# Launch vLLM-Omni server for IndexTTS2
#
# Usage from repository root:
# examples/online_serving/text_to_speech/indextts2/run_server.sh
# CUDA_VISIBLE_DEVICES=0 PORT=8092 MODEL=/path/to/IndexTeam/IndexTTS-2 examples/online_serving/text_to_speech/indextts2/run_server.sh
set -e
SCRIPT_DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)"
ROOT_DIR="$(cd -- "$SCRIPT_DIR/../../../.." && pwd)"
MODEL="${MODEL:-IndexTeam/IndexTTS-2}"
PORT="${PORT:-8092}"
DEPLOY_CONFIG="${DEPLOY_CONFIG:-$ROOT_DIR/vllm_omni/deploy/indextts2.yaml}"
echo "Starting IndexTTS2 server with model: $MODEL"
FLASHINFER_DISABLE_VERSION_CHECK=1 \
vllm serve "$MODEL" \
--host 0.0.0.0 \
--port "$PORT" \
--omni \
--trust-remote-code \
--deploy-config "$DEPLOY_CONFIG"