PMoS旨在创建一个深度学习一站式开发平台，使用可视化编程的方式完成模型构建、模型训练、模型部署全流程。

PMoS, which stands for Produce your Module at one Site. We strive to build a platform using visual programming that allows most people to enjoy the process of building deep-learning modules.

构建方案
Build Approaches

PMoS将分为多个模块来完成整体的构建，包括：

• 网络构建、网络训练支持模块flowing
• 其余模块正在计划中...

PMoS will be divided into multiple modules to complete the overall construction, including:

• network construction and network training support module flowing
• other modules are currently under planning...

模块说明
Modules

模块flowing
flowing Module

功能
Function

支持流程式定义网络，指定输入节点、算子节点、输出节点及其数据流向即可完成网络定义；支持流程式定义训练流程，指定每个步中的数据操作即可完成训练过程的定义。

flowing supports defining networks in a workflow-like manner. You just need to specify the input nodes, operator nodes, output nodes, and their data flow directions to complete the network definition. It also supports defining the training process in a similar way. Just specify the data operations in each step to complete the definition of the training process.

代码形式调用
Using by Code

./demo中可找到相关的示例，其中网络的定义示例项目为net_generate，网络的训练示例项目为segment。

You can find relevant examples in ./demo. The example project for network definition is net_generate, and the example project for network training is segment.

Web服务形式调用
Using by Web

启动方法
How to Boot

1. 访问flowing部署在 https://pmos.lohhhha.cn 的在线服务。
2. 按以下步骤本地启动flowing服务：
   1. 准备flowing的whl包，可以在本项目仓库下载，也可以自行下载源码打包。
   2. 在选定Python环境中安装flowing，推荐的Python版本为3.12，需要的依赖会自动获取。

      pip install flowing-xxxx.whl
      

   3. 启动flowing服务。

      python -m flowing.server.runner --port 54321 --host 127.0.0.1 --log_level INFO
      

   4. 根据启动参数，打开对应的Web服务，例如：127.0.0.1:54321。

1. Access the online service of flowing deployed at https://pmos.lohhhha.cn.
2. Follow these steps to start the flowing service locally:
   1. Prepare the whl package of flowing. You can download it from this repository or package it yourself by downloading the source code.
   2. Install flowing in the selected Python environment. The recommended Python version is 3.12, and the required dependencies will be obtained automatically.

      pip install flowing-xxxx.whl
      

   3. Start the flowing service.

      python -m flowing.server.runner --port 54321 --host 127.0.0.1 --log_level INFO
      

   4. Open the corresponding Web service according to the startup parameters, such as 127.0.0.1:54321.

使用方法
How to Use

flowing 使用说明

flowing Instruction