Why the "Launch Model" tab is empty under supervisor mode #4373
Description
System Info / 系統信息
`version: "3.8"
services:
xinference_supervisor:
image: xprobe/xinference:v1.14.0
container_name: xinference_supervisor
volumes:
- ./xinference_supervisor:/.xinference
ports:
- "9997:9997"
- "9996:9996"
environment:
# - XINFERENCE_HOME=/tmp/xinference
- XINFERENCE_MODEL_SRC=modelscope
command: xinference-supervisor -H 192.168.10.10 -p 9997 --supervisor-port 9996 --log-level debug
# command: xinference-local -H 0.0.0.0 -p 9997 --log-level debug
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]
restart: unless-stopped
`
Here is the docker compose file I haved been used. If I start the service as local, the "launch model" tab shows models and able to run model. However, when I start as supervisor, the tab become empty.
I do has a worker connected to the supervisor and the information is shown on the "Cluster information" tab.
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
v1.14.0
The command used to start Xinference / 用以启动 xinference 的命令
xinference-supervisor -H 10.16.249.249 -p 9997 --supervisor-port 9996 --log-level debug
Reproduction / 复现过程
Please replicate with my docker compose file
Expected behavior / 期待表现
The "Launch model" tab has models to select