README.md

Vega

中文

Introduction

Vega is an AutoML algorithm tool chain developed by Noah's Ark Laboratory, the main features are as follows:

1. Full pipeline capailities: The AutoML capabilities cover key functions such as Hyperparameter Optimization, Data Augmentation, Network Architecture Search (NAS), Model Compression, and Fully Train. These functions are highly decoupled and can be configured as required, construct a complete pipeline.
2. Industry-leading AutoML algorithms: provides Noah's Ark Laboratory's self-developed industry-leading algorithm and Model Zoo to download the State-of-the-art (SOTA) model.
3. High-concurrency neural network training capability: Provides high-performance trainers to accelerate model training and evaluation.

Algorithm list

Category	Algorithm	Description	reference
NAS	SM-NAS: Structural-to-Modular NAS	Two-stage object detection architecture search algorithm	Coming soon
NAS	CARS: Continuous Evolution for Efficient Neural Architecture Search	Structure Search Method of Multi-objective Efficient Neural Network Based on Continuous Evolution	ref
NAS	SR-EA	An Automatic Network Architecture Search Method for Super Resolution	ref
NAS	ESR-EA: Efficient Residual Dense Block Search for Image Super-resolution	Multi-objective image super-resolution based on network architecture search	ref
NAS	Adelaide-EA: SEGMENTATION-Adelaide-EA-NAS	Network Architecture Search Algorithm for Image Segmentation	ref
NAS	SP-NAS: Serial-to-Parallel Backbone Search for Object Detection	Serial-to-Parallel Backbone Search for Object Detection Efficient Search Algorithm for Object Detection and Semantic Segmentation in Trunk Network Architecture	ref
NAS	Auto-Lane: CurveLane-NAS	An End-to-End Framework Search Algorithm for Lane Lines	Coming soon
Model Compression	Quant-EA: Quantization based on Evolutionary Algorithm	Automatic mixed bit quantization algorithm, using evolutionary strategy to quantize each layer of the CNN network	ref
Model Compression	Prune-EA	Automatic channel pruning algorithm using evolutionary strategies	ref
HPO	ASHA: Asynchronous Successive Halving Algorithm	Dynamic continuous halving algorithm	ref
HPO	TPE: Tree-structured Parzen Estimator Approach	A hyperparameter optimization Algorithm Based on Tree - Structured Parzen Estimation	ref
HPO	BO: Bayesian Optimization	Bayesian optimization algorithm	ref
HPO	BOHB: Hyperband with Bayesian Optimization	Hyperband with Bayesian Optimization	ref
HPO	BOSS: Bayesian Optimization via Sub-Sampling	A universal hyperparameter optimization algorithm based on Bayesian optimization framework for resource-constraint hyper-parameters search	ref
Data Augmentation	PBA: Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules	Data augmentation based on PBT optimization	ref
Data Augmentation	cyclesr: CycleGAN + SR	Unsupervised style migration algorithm for low-level vision problem.	ref
Fully Train	Beyond Dropout: Feature Map Distortion to Regularize Deep Neural Networks	Neural network training (regularization) based on disturbance of feature map	ref
Fully Train	Circumventing Outliers of AutoAugment with Knowledge Distillation	Joint knowledge distillation and data augmentation for high performance classication model training, achieved 85.8% Top-1 accuracy on ImageNet 1k	Coming soon

Obtaining and Installing

The software has been released on the Releases page. Obtain the latest version and install it by following the instructions provided in the installation guide.

If you want to deploy Vega in local cluster, see the deployment guide.

Usage Guide

The Vega is highly modularized. You can configure the search space, search algorithm in a pipeline way. To run the Vega application is to load the configuration file and complete the AutoML process based on the configuration. Vega provides detailed operation examples for your reference. For details, see the examples. Example of running CARS algorithm:

cd examples
python3 ./run_example.py ./nas/cars/cars.yml

Therefore, before using the Vega, you need to fully understand the meaning of the configuration items. For details, see the Configuration Guide.

Note:

Before running an example, you need to configure the directory where the dataset and pre-trained models are located in the algorithm configuration file. Please refer to Example Reference.

Developer Guide

The Vega framework components are decoupled, and each functional component is combined using the registration mechanism to facilitate function and algorithm extension. For details about the Vega architecture and main mechanisms, see the Developer Guide.

In addition, you can refer to the Quick Start Guide to implement a simple network search function and quickly enter the Vega application development through practice.

During the development of the Vega application, the first problem is how to introduce the service data set to the Vega application. For details, see the Dataset Guide.

For different algorithms, you can refer doc Algorithm Development Guide. You can add the new algorithm to Vega step by step based on the example provided in this document.

In most Automl algorithms, the search space is closely related to the network. We try to unify the definition of the search space so that the same search space can adapt to different search algorithms. This is called the Fine-Grained Search Space Guide. Welcome to try it.

Of course, this document cannot solve all the problems. If you have any questions, please feel free to provide feedback through the issue. We will reply to you and solve your problems in a timely manner.

Reference List

object	refrence
User	Install Guide, Deployment Guide, Configuration Guide, Examples
Developer	API Reference, Developer Guide, Quick Start Guide, Dataset Guide, Algorithm Development Guide, Fine-Grained Search Space Guide

Cooperation and contribution

Welcome to use Vega. If you have any questions, ask for help, fix bugs, contribute algorithms, or improve documents, submit the issue in the community. We will reply to and communicate with you in a timely manner. Welcome to join our QQ chatroom (Chinese): 833345709.