ASR模型不能选择GPU部署 #4055
Description
System Info / 系統信息
Running Xinference with Docker? / 是否使用 Docker 运行 Xinfernece？
- docker / docker
- pip install / 通过 pip install 安装
- installation from source / 从源码安装
Version info / 版本信息
1.9.1
The command used to start Xinference / 用以启动 xinference 的命令
supervisor
services:
xinference-supervisor:
image: xprobe/xinference:v1.9.1-cpu
container_name: xinference-supervisor
network_mode: host
environment:
- NVIDIA_VISIBLE_DEVICES=all
- XINFERENCE_HOME=/data
- XINFERENCE_MODEL_SRC=modelscope
volumes:
- /root/data:/data
command: >
xinference-supervisor -H 192.168.1.33 --port 9997 --log-level DEBUG
worker
services:
xinference-worker:
image: xprobe/xinference:v1.9.1
container_name: xinference-worker
restart: always
network_mode: host
environment:
- XINFERENCE_HOME=${XINFERENCE_HOME}
- XINFERENCE_MODEL_SRC=${XINFERENCE_MODEL_SRC}
volumes:
- ${CACHE_PATH}:/data
command: >
xinference-worker -e "${ENDPOINT}" -H "${HOST}" --log-level ${LOG_LEVEL}
stdin_open: true
tty: true
deploy:
resources:
reservations:
devices:
- capabilities: [gpu]
docker-compose -f docker-compose.supervisor.yaml up -d
docker compose -f docker-compose.worker-gpu.yaml up -d
Reproduction / 复现过程
启动后，直接在webui去部署seaco-paraformer-zh
Expected behavior / 期待表现
可以选择GPU部署