Pipelines and Reuse with dasktl;dr: We use dask to accelerate parameter searches over machine learning pipelines by naming consistently. This work builds on work done with Andreas Mueller, Olivier Grisel, with suggestions by Gael Varoquaux Disclaimer: this blogpost discusses experimental and buggy code. Outline¶Computational costs of machine learning applications are multiplied by the following: Pipelines of several transform
Calculations with arrays bigger than your memory (dask arrays)
