Training-Free Multi-Objective and Many-Objective Evolutionary Neural Architecture Search with Synaptic Flow

An Vo, Tan Ngoc Pham, Van Bich Nguyen, Ngoc Hoang Luong

In SoICT 2022.

Setup

• Clone this repository
• Install packages

$ pip install -r requirements.txt

• Download NATS-bench, put it in the benchmark folder and follow instructions here

Usage

python search.py --method <method_name>
               --complexity_obj <complexity_objective> --pop_size <population_size>
               --n_gens <num_of_generations> --dataset <dataset_name> --seed <random_seed> 

Please see details in search.py.

# MOENAS
python search.py --method MOENAS --complexity_obj flops

# TF-MOENAS
python search.py --method TF-MOENAS --complexity_obj flops

# MaOENAS
python search.py --method MaOENAS

# TF-MOENAS
python search.py --method TF-MaOENAS 

Acknowledgement

Our source code is inspired by:

• pymoo: Multi-objective Optimization in Python
• NSGA-Net: Neural Architecture Search using Multi-Objective Genetic Algorithm
• Zero-Cost Proxies for Lightweight NAS
• NATS-Bench: Benchmarking NAS Algorithms for Architecture Topology and Size