cooka 0.1.1

A lightweight AutoML system.

Cooka

简体中文

Cooka is a lightweight and visualization toolkit to manage datasets and design model learning experiments through web UI. It using DeepTables and HyperGBM as experiment engine to complete feature engineering, neural architecture search and hyperparameter tuning automatically.

Features overview

Through the web UI provided by cooka you can:

• Add and analyze datasets
• Design experiment
• View experiment process and result
• Using models
• Export experiment to jupyter notebook

Screen shots：

The machine learning algorithms supported are ：

• XGBoost
• LightGBM
• Catboost

The neural networks supported are：

• WideDeep
• DeepFM
• xDeepFM
• AutoInt
• DCN
• FGCNN
• FiBiNet
• PNN
• AFM
• ...

The search algorithms supported are：

• Evolution
• MCTS(Monte Carlo Tree Search)
• ...

The supported feature engineering provided by scikit-learn and featuretools are：

• Scaler

  • StandardScaler
  • MinMaxScaler
  • RobustScaler
  • MaxAbsScaler
  • Normalizer

• Encoder

  • LabelEncoder
  • OneHotEncoder
  • OrdinalEncoder

• Discretizer

  • KBinsDiscretizer
  • Binarizer

• Dimension Reduction

  • PCA

• Feature derivation

  • featuretools

• Missing value filling

  • SimpleImputer

It can also extend the search space to support more feature engineering methods and modeling algorithms.

Installation

Using pip

The python version should be >= 3.6, for CentOS , install the system package:

pip install --upgrade pip
pip install cooka

Start the web server：

cooka server

Then open http://<your_ip:8000> with your browser to use cooka.

By default, the cooka configuration file is at ~/.config/cooka/cooka.py, to generate a template:

mkdir -p ~/.config/cooka/
cooka generate-config > ~/.config/cooka/cooka.py

Using Docker

Launch a Cooka docker container:

docker run -it -p 8000:8000 -p 8888:8888 -e NOTEBOOK_PORTAL="http://<your_ip>:8888"  datacanvas/cooka

Open http://<your_ip:8000> with your browser to visit cooka.

DataCanvas

Cooka is an open source project created by DataCanvas.