容器 | 镜像 | 组件关系 | 建议分配资源 |
bisheng-frontend | dataelement/bisheng-frontend | 应用程序: Nginx 软件依赖: Nginx 网络端口: 3001 | 1~2 cores 2~8G memory |
bisheng-backend | dataelement/bisheng-backend | 应用程序: /app/bisheng 软件依赖:langchain0.0.232, bisheng_langchain, fastapi, 系统依赖:Linux 环境 运行时:Python3.10 网络端口:7860 | 1~2 cores 2~8G memory |
bisheng-rt | dataelem/bisheng-rt | 应用程序: /opt/bisheng-rt/bin/rtserver f 软件依赖库:pytorch2.0.1, tensorflow 1.15.4, transformers 4.31.0 accelerate 0.21.0 系统依赖: Host安装Nvidia Driver Version: 515.76+ 运行时: Linux Binary环境和Python 3.8 文件系统:Host挂载的model repository目录,E.g., /opt/model-repository 网络端口: 9000, 9001, 9002 | 10~20 cores 64G~256G memory(内存) 24G+ GPU Memory GPU Memory由上线的模型数量决定,各模型单实例建议显存见:模型管理 128G Disk |
bisheng-unstructure | dataelem/bisheng-unstructure | 应用程序: /opt/bisheng-unstructured/bin/entrypoint.sh 软件依赖库:textlive 2023, pandoc, pymupdf, libreoffice 系统依赖: Ubuntu 20.04 运行时: Python 3.8 网络端口: 10010 | 2 cores 32G memory |
bisheng-mysql | mysql:5.7.40 | | 4 cores 4G memory 100G Disk |
bisheng-redis | redis:7.0.4 | | 2 cores 2G memory |
bisheng-etcd | quay.io/coreos/etcd:v3.5.5 | | 2 cores 2G memory |