Welcome to BenchENAS’s documentation!

Introduction

BenchENAS is a platform to help users conduct Evolutionary Algorithm based NAS (ENAS) research. BenchENAS has the following features:

1. Nine representative state-of-the-art ENAS algorithms have been implemented into BenchENAS. In addition, BenchENAS provides many popular data processing techniques, common image classification datasets, learning rate policices and optimizers for users to choose from.

2. The distributed training method designed for ENAS is implemented in BenchENAS to accelerate the fitness evaluation phase.

3. Good usability. All the ENAS algorithms in BenchENAS can be easily configured with different data settings and different trainer settings. And BenchENAS is implemented in python using very few third-party libraries for easy installation.

4. Easy extensibility. BenchENAS is easy to extend due to its highly modular design. Users can easily implement their ENAS algorithms in BenchENAS.

Contents:

• BenchENAS
  • init_compute
  • main
  • train package
  • comm package
  • compute package

Indices and tables

• Index

• Module Index

• Search Page