Auto-PyTorch

Finding the right architecture and hyperparameter settings for training a deep neural network is crucial to achieve top performance. Auto-PyTorch automates these two aspects by using multi-fidelity optimization and Bayesian optimization (BOHB) to search for the best settings.

The current version of Auto-PyTorch is an early alpha and only supports featurized data. The upcoming versions will also support image data, natural language processing, speech and videos.

References

  • Mendoza, Hector and Klein, Aaron and Feurer, Matthias and Springenberg, Jost Tobias and Urban, Matthias and Burkart, Michael and Dippel, Max and Lindauer, Marius and Hutter, Frank
    Towards Automatically-Tuned Deep Neural Networks
    In: AutoML: Methods, Sytems, Challenge