Overview

What is the DaCapo Toolbox?

The DaCapo Toolbox contains a set of convenient tools for training neural networks on large microscopy datasets. Specifically it focuses on providing a simple interface to many tools and techniques with an emphasis on common interfaces such as torch/numpy/dask and other common scientific computing packages. The main exception to this is the funlib.persistence.Array object. A thin wrapper around a dask array with added metadata necessary for reliable training on diverse microscopy datasets.