Skip to content

[ICLR 2021] "Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspective" by Wuyang Chen, Xinyu Gong, Zhangyang Wang

License

Notifications You must be signed in to change notification settings

VITA-Group/TENAS

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspective [PDF]

Language grade: Python MIT licensed

Wuyang Chen, Xinyu Gong, Zhangyang Wang

In ICLR 2021.

Overview

We present TE-NAS, the first published training-free neural architecture search method with extremely fast search speed (no gradient descent at all!) and high-quality performance.

Highlights:

  • Trainig-free and label-free NAS: we achieved extreme fast neural architecture search without a single gradient descent.
  • Bridging the theory-application gap: We identified two training-free indicators to rank the quality of deep networks: the condition number of their NTKs, and the number of linear regions in their input space.
  • SOTA: TE-NAS achieved extremely fast search speed (one 1080Ti, 20 minutes on NAS-Bench-201 space / four hours on DARTS space on ImageNet) and maintains competitive accuracy.

Prerequisites

  • Ubuntu 16.04
  • Python 3.6.9
  • CUDA 10.1 (lower versions may work but were not tested)
  • NVIDIA GPU + CuDNN v7.3

This repository has been tested on GTX 1080Ti. Configurations may need to be changed on different platforms.

Installation

  • Clone this repo:
git clone https://github.com/chenwydj/TENAS.git
cd TENAS
  • Install dependencies:
pip install -r requirements.txt

Usage

0. Prepare the dataset

  • Please follow the guideline here to prepare the CIFAR-10/100 and ImageNet dataset, and also the NAS-Bench-201 database.
  • Remember to properly set the TORCH_HOME and data_paths in the prune_launch.py.

1. Search

python prune_launch.py --space nas-bench-201 --dataset cifar10 --gpu 0
python prune_launch.py --space nas-bench-201 --dataset cifar100 --gpu 0
python prune_launch.py --space nas-bench-201 --dataset ImageNet16-120 --gpu 0
python prune_launch.py --space darts --dataset cifar10 --gpu 0
python prune_launch.py --space darts --dataset imagenet-1k --gpu 0

2. Evaluation

  • For architectures searched on nas-bench-201, the accuracies are immediately available at the end of search (from the console output).
  • For architectures searched on darts, please use DARTS_evaluation for training the searched architecture from scratch and evaluation.

Citation

@inproceedings{chen2020tenas,
  title={Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspective},
  author={Chen, Wuyang and Gong, Xinyu and Wang, Zhangyang},
  booktitle={International Conference on Learning Representations},
  year={2021}
}

Acknowledgement

About

[ICLR 2021] "Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspective" by Wuyang Chen, Xinyu Gong, Zhangyang Wang

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages