https://github.com/dataloop-ai/AutoML

ZazuML - easy AutoML for Object Detection
https://github.com/dataloop-ai/AutoML

automl

Last synced: 12 days ago
JSON representation

ZazuML - easy AutoML for Object Detection

• Host: GitHub
• URL: https://github.com/dataloop-ai/AutoML
• Owner: dataloop-ai
• License: mit
• Created: 2019-12-26T12:23:21.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2023-05-22T22:19:39.000Z (about 1 year ago)
• Last Synced: 2024-06-18T06:15:41.296Z (18 days ago)
• Topics: automl
• Language: Python
• Homepage:
• Size: 34.9 MB
• Stars: 336
• Watchers: 23
• Forks: 44
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• awesome-automl - dataloop-ai/zazuml - source AutoML framework for object detection. Currently this project contains a model & hyper-parameter tuner, auto augmentations, trial manager and prediction trigger, already loaded with your top preforming model-checkpoint. A working pipeline ready to be plugged into your product, simple as that (Libraries)

README

        
![Logo](./images/ZazuML.jpeg)



   

This is an easy to use open-source AutoML framework for object detection. This project contains an

architecture search + hyper-parameter search + augmentations search, as well as trial manager. 

Create an ML Pipeline taylor made for simple integration into your project or code.







### *Algorithms*

| *Augmentations Search* | *Hyper-Parameter Search* | *Architecture Search*

| :----:         |     :----:      |    :----:      |

|  
Fast-Autoaugment
    | 


Random Search

HyperBand

     | 
Your Model
  |

### *Super EASY to use!*


![](./images/new_zazu_tool.gif)  

## *Why ZazuML?*

You might be building an ML pipeline to avoid model performance degeneration, or maybe you're just too busy to download, 

debug, and tune your own model. Either way, you shouldn't be focusing your efforts on simple things like detection. There is

a entire world out there waiting for you to explore, try your hand at trajectory prediction or action recognition and let *ZazuML*

free you up from all the boring, tedious stuff.

- Get started with [Quick Start](DOCS/GETTINGSTARTED.md)

- Read up on [Configuring ZazuML](DOCS/CONFIGURINGZAZU.md)

- Launch [ZazuML remotely](DOCS/REMOTEZAZU.md)

- Take a look [Under The Hood](DOCS/UNDERTHEHOOD.md)

- Customize by [adding your own models](DOCS/ADDINGMODELS.md)

## *TO DO*

- Increase search space.

- NAS to replace some of the HP search.

- Intelligent Losses to replace some of the HP search.

## *Want to contribute*

If you're interested in becoming a collaborator or just have some questions, feel free to join our slack community.

## *References*

Some of the code was inspired by [keras-tuner](https://github.com/keras-team/keras-tuner)