Auto-Keras is a highly efficient AutoML system built on the basis of Keras. DATA Lab at Texas A&M University developed it to make machine learning easily accessible to everyone.

The platform is an open-source library the offers a simple and effective approach for finding top-performing models. These models are used for a variety of predictive modeling tasks, such as classification and regression datasets.

AutoKeras is an AutoML implementation for deep learning that leverages neural architecture search. It utilizes Neural Architecture Search along with Bayesian optimization for guiding Bayesian optimization network morphism for highly efficient neural network search.

Project Background

• • Library: Auto-Keras AutoML platform 
  • Author: Haifeng Jin, François Chollet, Qingquan Song
  • Initial Release: 2008
  • Type: Works with deep learning, classification, and regression
  • License: Apache-2.0 License
  • Contains: ImageClassifier, ImageRegressor, TextClassifier, TextRegressor, StructuredDataClassifier, Node, and Base Class
  • Language: Python
  • GitHub: /keras-team/autokeras
  • Runs On: Linux
  • GitHub Discussions: /keras-team/autokeras/discussions
  • Stackflow: /tags/auto-keras

Applications

• Image Classification
• Image Regression
• Text Classification
• Text Regression
• Structured Data Classification
• Structured Data Regression 
• TimeSeriesForecaster