Welcome to Easy Lightning’s documentation!

Introduction

Welcome to the Easy Lightning documentation. This site provides comprehensive guides and API references for building and configuring your Easy Lightning models. Use the table of contents below to navigate through the modules and configuration settings.

EasyLightning is not limited to the specific datasets, models, or experimental setups presented in repository.

• New datasets can be integrated simply by converting them into the standardized format used by the framework.

• Models can be added easily as PyTorch nn.Module classes, with minimal configuration to define their inputs and outputs.

• The experimental setup is fully configurable via a YAML file, allowing users to explore custom scenarios beyond those covered here.

• The framework also supports further extensibility — including custom loss functions, optimizers, and more.

So far, we’ve developed EasyTorch and EasyRec. If you’d like to help us develop more modules or if you find any bugs, feel free to get in touch with us!

Get Started

• Installation
• Quick Start Guide

User Guide

• Config
• Easy_Rec Config
• Easy_Torch Config

Developer Guide

• Custom Metrics
  • 1. How to Define Metrics in model.yaml
  • 2. How to Load Metrics in Your Code
  • 3. Using the Metrics in the Standard Pipeline
  • Additional Notes
• Custom Loss
  • 1. How to Define a Loss in model.yaml
  • 2. How to Load the Loss in Your Code
  • 3. Using the Loss in the Standard Pipeline
  • Additional Notes
• Custom Dataset
• Custom Callbacks
  • 1. Folder Structure
  • 2. Define Your Callback
  • 3. Package Initialization
  • 4. Make the Package Importable
  • 5. Reference the Callback in Your Config
  • 6. Initialize the Trainer

API Reference

• Easy_Rec package
• Easy_Torch package