Deep Learning with Apache Spark and TensorFlow The Databricks Blog

The TensorFlow library automates the advent of schooling algorithms for neural networks of lots of shapes and sizes. The actual technique of constructing a neural community, even though, is more complex than simply working some characteristic on a dataset. There are typically a number of very crucial hyperparameters configuration parameters in layman’s terms to set, which influences how the model is expert. Picking the correct parameters ends up in high performance, while bad parameters can lead to prolonged schooling and bad performance.

In observe, desktop studying practitioners rerun an analogous model distinctive times with alternative hyperparameters so as find one of the best set. This is a classical method called hyperparameter tuning.

See also  AI Weirdness