Welcome to Training Noodles documentation!

Training Noodles is a simple and powerful tool to help training multiple programs on multiple servers.

Links

• GitHub repository

• PyPI package page

Introduction

• 1. Introduction
  • 1.1. What is Noodles
  • 1.2. Why Created This Tool
  • 1.3. Features
• 2. Use Case
  • 2.1. Train 4 Experiments on 3 Servers
• 3. How Noodles Works
  • 3.1. Deploy Experiments to Servers
  • 3.2. Check Requirements
  • 3.3. Deploy One Experiment

User Guide

• 1. Installation
  • 1.1. Prerequisites
  • 1.2. Install Noodles via PyPI
• 2. Linux-based Terminal for Windows
  • 2.1. Install and Uninstall Packages
  • 2.2. Use git-sdk in Visual Studio Code
• 3. Usage
  • 3.1. Run Noodles
  • 3.2. Choose Only Some Experiments
  • 3.3. Command Types
  • 3.4. Spec File
  • 3.5. Debug
• 4. Write a Minimalist Spec File
  • 4.1. Create a File
  • 4.2. Add a Name
  • 4.3. Add a Description
  • 4.4. Add Experiments
  • 4.5. Add Servers
  • 4.6. Final Spec File
  • 4.7. Run the Spec File
• 5. See Other Examples
  • 5.1. Two Locals
  • 5.2. Two Servers with Same User
  • 5.3. Error Handling
  • 5.4. File Transfer
  • 5.5. Local Tmux
  • 5.6. Long Process
  • 5.7. Train TensorFlow Examples
• 6. Caveats
  • 6.1. Set Working Directory
  • 6.2. Set Environment Variables
  • 6.3. Choose the Type of Shells

References

• 1. Command Line Arguments
  • 1.1. Usage
  • 1.2. Positional Arguments
  • 1.3. Optional Arguments
• 2. Spec Options
  • 2.1. General Options
  • 2.2. Experiments
  • 2.3. Servers
  • 2.4. Requirements
  • 2.5. Deployment
  • 2.6. Error Handling
  • 2.7. Shell Commands
• 3. Environment Variables
  • 3.1. Environment Variables in Spec
  • 3.2. Extra Environment Variables

Indices and tables

• Index

• Search Page