ML Example Output writes a set of labeled examples to a file to directly use with ML Train Regression. ML Example Output expects an input geometry consisting exclusively of packed primitives, each of which represents a single labeled example. See ML Example for more details about representing examples using geometry.

The data stored in the attributes and volumes specified in the multiparms are written out in linear order to disk. When writing python snippets in ML Train regression, for example using convolution layers, It is important to know in what order the floating-point values stored in a volume are written out. This becomes relevant when reshaping flat tensors to make them suitable as inputs to a convolution layer. For example, if we have a volume that’s an RGB image, it is useful to know that the first non-batch dimension would be the color, followed by the x dimension and the y dimension. In general for tuple-valued volumes, the first nonbatch dimension would correspond to the tuple index, while the dimensions after that would correspond to indices into voxels.

This node can be referenced from TOPs using a ROP Fetch to make writing out the data set part of a recipe for doing ML.

See Machine Learning documentation for more general information.